Polysubstance and tr. completion
Analyze the association between polysubstance at admission and tr. compleiton longitudinally along treatments, accounting for irregular observations.
Code
rm(list = ls())
unlink("proposal_grant_23_24_files", recursive=T)
#fuentes:
#https://rpubs.com/georgy_makarov/897844
path<-paste0(getwd(),'/env')
#Sys.setenv(RETICULATE_PYTHON = "")
#Sys.setenv(RETICULATE_PYTHON = Sys.which("python"))
#reticulate::py_config()
#use_python(paste0(path,"/Scripts/python.exe"))
#Sys.setenv(LD_LIBRARY_PATH = paste0(path,"/Lib"))
#Sys.setenv(LD_LIBRARY_PATH_64 = paste0(path,"/Lib"))
#instalar paquetes de funcionalidades básicas para tener ubicaciones relativas y acceso a python (reticulate)
if(!require(reticulate)){install.packages("reticulate")}Loading required package: reticulateWarning: package 'reticulate' was built under R version 4.1.3Code
if(!require(rstudioapi)){install.packages("rstudioapi")}Loading required package: rstudioapiWarning: package 'rstudioapi' was built under R version 4.1.3Code
invisible("Create env")
#https://stackoverflow.com/questions/54043607/how-to-set-pyenv-python-for-reticulate
#Directory H:/Mi unidad/PERSONAL ANDRES/UCH_salud_publica/asignaturas/env is not a Python virtualenv
#virtualenv_create(envname = path, packages = c("pip", "statsmodels", "matplotlib", "numpy", "pandas", "scipy"))
# "C:/Users/andre/anaconda3/python.exe" -m venv "H:/Mi unidad/PERSONAL ANDRES/UCH_salud_publica/asignaturas/9_Computacion_Estadistica/env"
#FUENTES:
#https://rstudio.github.io/reticulate/articles/versions.html
#Virtual environment functions are not supported on Windows (the use of conda environments is recommended on Windows).
invisible("Use environment")
#https://ugoproto.github.io/ugo_r_doc/pdf/reticulate.pdf
# tx <- readLines(paste0(path,"/pyvenv.cfg"))
# tx[[1]] <- paste0("home = ",gsub('/', '\\', paste0(path,"/Scripts/python.exe"), fixed=T))
# tx[[3]] <- "version = 3.8.0"
#writeLines(tx, con=paste0(path,"/pyvenv.cfg"))
#H:/Mi unidad/PERSONAL ANDRES/UCH_salud_publica/asignaturas/env/Scripts/python.exe"
#use_virtualenv(path)
#usar entorno virtual ya creado
#información sobre entorno virtual
#py_discover_config()
#conda_python(envname = "r-scrublet")
# FUENTES
#https://akrabat.com/creating-virtual-environments-with-pyenv/
#https://rstudio.github.io/reticulate/reference/install_python.html
#https://github.com/pyenv/pyenv/wiki#suggested-build-environment
#https://github.com/pyenv/pyenv
#https://stackoverflow.com/questions/56755156/reticulate-not-setting-python-path
#https://github.com/rstudio/reticulate/issues/291#issuecomment-437143751
#https://github.com/pyenv/pyenv
#https://github.com/pyenv-win/pyenv-win#installation
#https://stackoverflow.com/questions/52060867/how-to-use-pip-for-pyenv
#https://github.com/pyenv/pyenv/issues/2417Code
!pyenv install -l | findstr 3.8
!pip install --upgrade pyenv-win
!env PYTHON_CONFIGURE_OPTS="--enable-shared" pyenv install
!env PYTHON_CONFIGURE_OPTS="--enable-shared" pyenv install 3.7.5
!pyenv buildCode
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
# `r format(Sys.time(),'%B %d, %Y')`
# Data import
load("an_grant_23_24.RData")
# List all of the objects names in RData:
#ls(.GlobalEnv)
#ls(new_environment)
#rm(new_environment)
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_Packages
Database consolidation
We used Base_fiscalia_v15f_grant_23_24 instead of v16, because we wanted to capture more treatments from each people.
Code
invisible("Label variables")
#2023-07-19
attr(Base_fiscalia_v15f_grant_23_24$motivodeegreso_mod_imp_rec,"label") <- "Complete status of treatment (binary)"
attr(Base_fiscalia_v15f_grant_23_24$offense_after_adm,"label") <- "Committing an offense after admission (binary)"
attr(Base_fiscalia_v15f_grant_23_24$age_at_censor_date,"label") <- "Age at censorship"
attr(Base_fiscalia_v15f_grant_23_24$age_offending_imp,"label") <- "Age at offending"
attr(Base_fiscalia_v15f_grant_23_24$age_tr_comp_imp,"label") <- "Age at completing tr."
#attr(Base_fiscalia_v15f_grant_23_24$tr_modality,"label") <- "Treatment Modality"
attr(Base_fiscalia_v15f_grant_23_24$time_to_off_from_adm,"label") <- "Time to offense from admission"
attr(Base_fiscalia_v15f_grant_23_24$time_to_drop_from_adm,"label") <- "Time to dropout from admission"
#2023-07-31
attr(Base_fiscalia_v15f_grant_23_24$age_at_death,"label") <- "Age at death"
attr(Base_fiscalia_v15f_grant_23_24$date_death,"label") <- "Time to death from admission"
attr(Base_fiscalia_v15f_grant_23_24$event_death,"label") <- "Death from admission"
attr(Base_fiscalia_v15f_grant_23_24$event_comp,"label") <- "Event: tr.completion"
attr(Base_fiscalia_v15f_grant_23_24$event_offense,"label") <- "Event: offense"
attr(Base_fiscalia_v15f_grant_23_24$edad_al_ing_1,"label") <- "Age (admission to treatment)"
attr(Base_fiscalia_v15f_grant_23_24$sex,"label") <- "Sex"
attr(Base_fiscalia_v15f_grant_23_24$edad_ini_cons,"label") <- "Age of Onset of Substance Use"
attr(Base_fiscalia_v15f_grant_23_24$escolaridad_rec,"label") <- "Educational Attainment"
attr(Base_fiscalia_v15f_grant_23_24$sus_principal_mod,"label") <- "Primary Substance (admission to treatment)"
attr(Base_fiscalia_v15f_grant_23_24$freq_cons_sus_prin,"label") <- "Frequency of Substance Use (Primary Substance)"
attr(Base_fiscalia_v15f_grant_23_24$condicion_ocupacional_corr,"label") <- "Corrected Occupational Status (f)"
attr(Base_fiscalia_v15f_grant_23_24$policonsumo,"label") <- "Co-occurring Substance Use Disorders (Polysubstance use)"
#20203-07-17
attr(Base_fiscalia_v15f_grant_23_24$otras_sus1_mod,"label") <- "First additional substance at admission"
attr(Base_fiscalia_v15f_grant_23_24$otras_sus2_mod,"label") <- "Second additional substance at admission"
attr(Base_fiscalia_v15f_grant_23_24$otras_sus3_mod,"label") <- "Third additional substance at admission"
attr(Base_fiscalia_v15f_grant_23_24$num_hijos_mod_joel_bin,"label") <- "Number of Children (dichotomized)"
attr(Base_fiscalia_v15f_grant_23_24$tenencia_de_la_vivienda_mod,"label") <- "Housing Situation (Tenure Status)"
attr(Base_fiscalia_v15f_grant_23_24$macrozona,"label") <- "Macro Administrative Zone in Chile"
attr(Base_fiscalia_v15f_grant_23_24$n_off_vio,"label") <- "Violent Criminal Offenses (Pre-Treatment)"
attr(Base_fiscalia_v15f_grant_23_24$n_off_acq,"label") <- "Acquisitive Criminal Offenses (Pre-Treatment)"
attr(Base_fiscalia_v15f_grant_23_24$n_off_sud,"label") <- "Substance-Related Criminal Offenses (Pre-Treatment)"
attr(Base_fiscalia_v15f_grant_23_24$n_off_oth,"label") <- "Other Criminal Offenses (Pre-Treatment)"
attr(Base_fiscalia_v15f_grant_23_24$dg_cie_10_rec,"label") <- "Psychiatric Comorbidity (ICD-10)"
attr(Base_fiscalia_v15f_grant_23_24$dg_trs_cons_sus_or,"label") <- "SUD Severity (Dependence status)"
attr(Base_fiscalia_v15f_grant_23_24$clas_r,"label") <- "Urbanicity"
attr(Base_fiscalia_v15f_grant_23_24$porc_pobr,"label") <- "Percentage of people in poverty"
attr(Base_fiscalia_v15f_grant_23_24$sus_ini_mod_mvv,"label") <- "Primary Substance (initial diagnosis)"
attr(Base_fiscalia_v15f_grant_23_24$ano_nac_corr,"label") <- "Corrected birth year"
attr(Base_fiscalia_v15f_grant_23_24$con_quien_vive_joel,"label") <- "Cohabitation status (Recoded) (f)"
attr(Base_fiscalia_v15f_grant_23_24$fis_comorbidity_icd_10,"label") <- "Physical Comorbidity (ICD-10)"
attr(Base_fiscalia_v15f_grant_23_24$n_post_off_vio,"label") <- "Count of Violent Criminal Offenses (Post-Treatment)"
attr(Base_fiscalia_v15f_grant_23_24$n_post_off_acq,"label") <- "Count of Acquisitive Criminal Offenses (Post-Treatment)"
attr(Base_fiscalia_v15f_grant_23_24$n_post_off_sud,"label") <- "Count of Substance-Related Criminal Offenses (Post-Treatment)"
attr(Base_fiscalia_v15f_grant_23_24$n_post_off_oth,"label") <- "Count of Other Criminal Offenses (Post-Treatment)"
attr(Base_fiscalia_v15f_grant_23_24$n_post_off,"label") <- "Count of Post-treatment Offenses"
#2024-01-27
attr(Base_fiscalia_v15f_grant_23_24$compromiso_biopsicosocial,"label") <- "Biopsychosocial compromise"
attr(Base_fiscalia_v15f_grant_23_24$origen_ingreso_mod,"label") <- "Treatment Admission Motive"
Base_fiscalia_v15f_grant_23_24$condicion_ocupacional_corr24 <-
dplyr::if_else(Base_fiscalia_v15f_grant_23_24$condicion_ocupacional_corr=="Looking for a job for the first time","Unemployed", Base_fiscalia_v15f_grant_23_24$condicion_ocupacional_corr, missing=NA_character_)
Base_fiscalia_v15f_grant_23_24$condicion_ocupacional_corr24 <-
dplyr::if_else(Base_fiscalia_v15f_grant_23_24$condicion_ocupacional_corr=="No activity","Inactive", Base_fiscalia_v15f_grant_23_24$condicion_ocupacional_corr24, missing=NA_character_)
Base_fiscalia_v15f_grant_23_24$condicion_ocupacional_corr24 <-
dplyr::if_else(Base_fiscalia_v15f_grant_23_24$condicion_ocupacional_corr=="Not seeking for work","Inactive", Base_fiscalia_v15f_grant_23_24$condicion_ocupacional_corr24, missing=NA_character_)
#Si la persona No Trabajó, interesa saber si está Desocupada (“busca trabajo por primera vez” o “cesante”) o
#Inactiva (que corresponde al resto de la opciones, quehaceres del hogar, estudiando, jubilado, etc.)
Base_fiscalia_v15f_grant_23_24 <- Base_fiscalia_v15f_grant_23_24 %>%
dplyr::mutate(freq_cons_sus_prin_ord = dplyr::case_when(
freq_cons_sus_prin == "Less than 1 day a week" ~ "1. Less than 1 day a week",
freq_cons_sus_prin == "1 day a week or more" ~ "2. 1 day a week or more",
freq_cons_sus_prin == "2 to 3 days a week" ~ "3. 2 to 3 days a week",
freq_cons_sus_prin == "4 to 6 days a week" ~ "4. 4 to 6 days a week",
freq_cons_sus_prin == "Daily" ~ "5. Daily"
)) %>%
dplyr::mutate(clas_r= dplyr::case_when(clas_r=="Mixta"~"Mixed", clas_r=="Urbana"~"Urban",clas_r=="Rural"~"Rural", T~clas_r)) %>%
dplyr::mutate(tenencia_de_la_vivienda_mod = dplyr::case_when(clas_r=="Illegal Settlement"~"Others", T~tenencia_de_la_vivienda_mod)) %>%
#2024-04-30: Divide type of plan according to program
dplyr::mutate(tipo_de_plan_2_mod= dplyr::case_when(as.character(tipo_de_plan_2)=="M-PR"~"WO residential", tipo_de_plan_2=="PG-PR"~ "GP residential",tipo_de_plan_2=="M-PAI"~ "WO intensive ambulatory",tipo_de_plan_2=="PG-PAI"~ "GP intensive ambulatory",grepl("PAB",tipo_de_plan_2)~"basic ambulatory", T~NA_character_)) %>%
dplyr::mutate(tipo_de_plan_2= dplyr::case_when(grepl("PR",tipo_de_plan_2)~"residential",grepl("PAB",tipo_de_plan_2)~"basic ambulatory",grepl("PAI",tipo_de_plan_2)~"intensive ambulatory", T~NA_character_))
#%>% #janitor::tabyl(tipo_de_plan_2_mod)
attr(Base_fiscalia_v15f_grant_23_24$condicion_ocupacional_corr24,"label") <- "Corrected Occupational Status"
attr(Base_fiscalia_v15f_grant_23_24$freq_cons_sus_prin_ord,"label") <- "Frequency of Substance Use (Primary Substance)"
# Check the new variable
table(Base_fiscalia_v15f_grant_23_24$freq_cons_sus_prin_ord)
1. Less than 1 day a week 2. 1 day a week or more 3. 2 to 3 days a week
4178 5610 23716
4. 4 to 6 days a week 5. Daily
13923 37200 Code
vars_cov <- c("motivodeegreso_mod_imp_rec", "edad_al_ing_1", "sex", "edad_ini_cons", "escolaridad_rec", "sus_principal_mod", "freq_cons_sus_prin_ord", "condicion_ocupacional_corr24", "policonsumo", "num_hijos_mod_joel_bin", "tenencia_de_la_vivienda_mod", "macrozona", "dg_cie_10_rec", "dg_trs_cons_sus_or", "clas_r", "porc_pobr", "sus_ini_mod_mvv", "ano_nac_corr", "con_quien_vive_joel", "fis_comorbidity_icd_10", "compromiso_biopsicosocial", "origen_ingreso_mod", "tipo_centro_pub","tipo_de_plan_2", "tipo_de_plan_2_mod")
#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Function to format CreateTableOne into a database")
as.data.frame.TableOne <- function(x, ...) {capture.output(print(x,
showAllLevels = TRUE, varLabels = T,...) -> x)
y <- as.data.frame(x)
y$characteristic <- dplyr::na_if(rownames(x), "")
y <- y %>%
fill(characteristic, .direction = "down") %>%
dplyr::select(characteristic, everything())
rownames(y) <- NULL
y}
#_#_#_#_#_#_#_#_#_#_#_#_#_
tbone_desc_merge_total_grant_23_24<-
CreateTableOne(vars=c(setdiff(vars_cov, "policonsumo"), "offense_after_adm"), data= Base_fiscalia_v15f_grant_23_24[,c("motivodeegreso_mod_imp_rec", "edad_al_ing_1", "sex", "edad_ini_cons",
"escolaridad_rec", "sus_principal_mod", "freq_cons_sus_prin_ord",
"condicion_ocupacional_corr24", "policonsumo", "num_hijos_mod_joel_bin",
"tenencia_de_la_vivienda_mod", "macrozona", "dg_cie_10_rec", "dg_trs_cons_sus_or",
"clas_r", "porc_pobr", "sus_ini_mod_mvv", "ano_nac_corr", "con_quien_vive_joel",
"fis_comorbidity_icd_10", "compromiso_biopsicosocial", "origen_ingreso_mod","tipo_centro_pub")], factorVars = setdiff(vars_cov, c("motivodeegreso_mod_imp_rec","edad_al_ing_1", "edad_ini_cons","ano_nac_corr", "porc_pobr")), smd=T, strata="policonsumo", addOverall = T, includeNA=T, test=T)Warning in ModuleReturnVarsExist(vars, data): The data frame does not have: tipo_de_plan_2 tipo_de_plan_2_mod offense_after_adm Dropped
Warning in ModuleReturnVarsExist(factorVars, data): The data frame does not have: tipo_de_plan_2 tipo_de_plan_2_mod Dropped
We compared the characteristics of people who used multiple substances and those who did not in the total database between 2013 and 2019.
Code
as.data.frame.TableOne(tbone_desc_merge_total_grant_23_24, smd=T, nonnormal= T)%>%
dplyr::mutate(char2=characteristic) %>%
tidyr::fill(char2) %>%
dplyr::select(char2,everything()) %>%
dplyr::mutate(level=ifelse(is.na(level),"[Missing]",level)) %>%
dplyr::mutate(char2=dplyr::case_when(characteristic=="NA"~NA_character_,TRUE~as.character(characteristic))) %>%
format_cells(1, 1:length(names(.)), "bold") %>%
dplyr::select(-1) %>%
dplyr::mutate(across(.fns = ~ str_replace_all(., pattern = "\\( ", replacement = "\\("))) %>%
dplyr::mutate(across(.fns = ~ str_trim(.))) %>%
knitr::kable(size=10, format="markdown",caption= "Summary descriptives, Polysubstance(1) and no Polysubstance use (0)", escape=T)| characteristic | level | Overall | 0 | 1 | p | test | SMD |
|---|---|---|---|---|---|---|---|
| n | 85048 | 22552 | 62496 | ||||
| Complete status of treatment (binary) (%) | Treatment completion | 19278 (22.7) | 6404 (28.4) | 12874 (20.6) | <0.001 | 0.183 | |
| Complete status of treatment (binary) (%) | Treatment non-completion | 65761 (77.3) | 16148 (71.6) | 49613 (79.4) | |||
| Complete status of treatment (binary) (%) | [Missing] | 9 (0.0) | 0 (0.0) | 9 (0.0) | |||
| Age (admission to treatment) (median [IQR]) | 34.21 [27.41, 43.17] | 41.71 [32.29, 51.21] | 32.33 [26.45, 39.90] | <0.001 | nonnorm | 0.743 | |
| Sex (%) | Men | 64028 (75.3) | 16502 (73.2) | 47526 (76.0) | <0.001 | 0.066 | |
| Sex (%) | Women | 21020 (24.7) | 6050 (26.8) | 14970 (24.0) | |||
| Age of Onset of Substance Use (median [IQR]) | 15.00 [14.00, 18.00] | 17.00 [14.00, 20.00] | 15.00 [13.00, 17.00] | <0.001 | nonnorm | 0.425 | |
| Educational Attainment (%) | 1-More than high school | 14068 (16.5) | 3215 (14.3) | 10853 (17.4) | <0.001 | 0.230 | |
| Educational Attainment (%) | 2-Completed high school or less | 46653 (54.9) | 11143 (49.4) | 35510 (56.8) | |||
| Educational Attainment (%) | 3-Completed primary school or less | 23943 (28.2) | 8044 (35.7) | 15899 (25.4) | |||
| Educational Attainment (%) | [Missing] | 384 (0.5) | 150 (0.7) | 234 (0.4) | |||
| Primary Substance (admission to treatment) (%) | Alcohol | 28859 (33.9) | 14212 (63.0) | 14647 (23.4) | <0.001 | 0.890 | |
| Primary Substance (admission to treatment) (%) | Cocaine hydrochloride | 16151 (19.0) | 2337 (10.4) | 13814 (22.1) | |||
| Primary Substance (admission to treatment) (%) | Cocaine paste | 32681 (38.4) | 4511 (20.0) | 28170 (45.1) | |||
| Primary Substance (admission to treatment) (%) | Marijuana | 5771 (6.8) | 972 (4.3) | 4799 (7.7) | |||
| Primary Substance (admission to treatment) (%) | Other | 1585 (1.9) | 519 (2.3) | 1066 (1.7) | |||
| Primary Substance (admission to treatment) (%) | [Missing] | 1 (0.0) | 1 (0.0) | 0 (0.0) | |||
| Frequency of Substance Use (Primary Substance) (%) | 1. Less than 1 day a week | 4178 (4.9) | 1530 (6.8) | 2648 (4.2) | <0.001 | 0.186 | |
| Frequency of Substance Use (Primary Substance) (%) | 2. 1 day a week or more | 5610 (6.6) | 1922 (8.5) | 3688 (5.9) | |||
| Frequency of Substance Use (Primary Substance) (%) | 3. 2 to 3 days a week | 23716 (27.9) | 6641 (29.4) | 17075 (27.3) | |||
| Frequency of Substance Use (Primary Substance) (%) | 4. 4 to 6 days a week | 13923 (16.4) | 3580 (15.9) | 10343 (16.5) | |||
| Frequency of Substance Use (Primary Substance) (%) | 5. Daily | 37200 (43.7) | 8744 (38.8) | 28456 (45.5) | |||
| Frequency of Substance Use (Primary Substance) (%) | [Missing] | 421 (0.5) | 135 (0.6) | 286 (0.5) | |||
| Corrected Occupational Status (%) | Employed | 41407 (48.7) | 12657 (56.1) | 28750 (46.0) | <0.001 | 0.276 | |
| Corrected Occupational Status (%) | Inactive | 14258 (16.8) | 4206 (18.7) | 10052 (16.1) | |||
| Corrected Occupational Status (%) | Unemployed | 29382 (34.5) | 5688 (25.2) | 23694 (37.9) | |||
| Corrected Occupational Status (%) | [Missing] | 1 (0.0) | 1 (0.0) | 0 (0.0) | |||
| Number of Children (dichotomized) (%) | 0 | 20138 (23.7) | 4161 (18.5) | 15977 (25.6) | <0.001 | 0.172 | |
| Number of Children (dichotomized) (%) | 1 | 64165 (75.4) | 18191 (80.7) | 45974 (73.6) | |||
| Number of Children (dichotomized) (%) | [Missing] | 745 (0.9) | 200 (0.9) | 545 (0.9) | |||
| tenencia_de_la_vivienda_mod (%) | Illegal Settlement | 906 (1.1) | 200 (0.9) | 706 (1.1) | <0.001 | 0.267 | |
| tenencia_de_la_vivienda_mod (%) | Others | 2354 (2.8) | 609 (2.7) | 1745 (2.8) | |||
| tenencia_de_la_vivienda_mod (%) | Owner/Transferred dwellings/Pays Dividends | 29934 (35.2) | 9863 (43.7) | 20071 (32.1) | |||
| tenencia_de_la_vivienda_mod (%) | Renting | 14566 (17.1) | 3835 (17.0) | 10731 (17.2) | |||
| tenencia_de_la_vivienda_mod (%) | Stays temporarily with a relative | 32609 (38.3) | 6829 (30.3) | 25780 (41.3) | |||
| tenencia_de_la_vivienda_mod (%) | [Missing] | 4679 (5.5) | 1216 (5.4) | 3463 (5.5) | |||
| Macro Administrative Zone in Chile (%) | Center | 64341 (75.7) | 16681 (74.0) | 47660 (76.3) | <0.001 | 0.293 | |
| Macro Administrative Zone in Chile (%) | North | 12151 (14.3) | 2243 (9.9) | 9908 (15.9) | |||
| Macro Administrative Zone in Chile (%) | South | 8536 (10.0) | 3625 (16.1) | 4911 (7.9) | |||
| Macro Administrative Zone in Chile (%) | [Missing] | 20 (0.0) | 3 (0.0) | 17 (0.0) | |||
| Psychiatric Comorbidity (ICD-10) (%) | Without psychiatric comorbidity | 32337 (38.0) | 10022 (44.4) | 22315 (35.7) | <0.001 | 0.184 | |
| Psychiatric Comorbidity (ICD-10) (%) | Diagnosis unknown (under study) | 16099 (18.9) | 3575 (15.9) | 12524 (20.0) | |||
| Psychiatric Comorbidity (ICD-10) (%) | With psychiatric comorbidity | 36612 (43.0) | 8955 (39.7) | 27657 (44.3) | |||
| SUD Severity (Dependence status) (%) | Drug dependence | 61836 (72.7) | 13974 (62.0) | 47862 (76.6) | <0.001 | 0.321 | |
| SUD Severity (Dependence status) (%) | Hazardous consumption | 23211 (27.3) | 8577 (38.0) | 14634 (23.4) | |||
| SUD Severity (Dependence status) (%) | [Missing] | 1 (0.0) | 1 (0.0) | 0 (0.0) | |||
| clas_r (%) | Mixed | 8205 (9.6) | 2850 (12.6) | 5355 (8.6) | <0.001 | 0.272 | |
| clas_r (%) | Rural | 7148 (8.4) | 2977 (13.2) | 4171 (6.7) | |||
| clas_r (%) | Urban | 69693 (81.9) | 16724 (74.2) | 52969 (84.8) | |||
| clas_r (%) | [Missing] | 2 (0.0) | 1 (0.0) | 1 (0.0) | |||
| Percentage of people in poverty (median [IQR]) | 0.11 [0.07, 0.16] | 0.11 [0.07, 0.15] | 0.12 [0.08, 0.17] | <0.001 | nonnorm | 0.104 | |
| Primary Substance (initial diagnosis) (%) | Alcohol | 46623 (54.8) | 15574 (69.1) | 31049 (49.7) | <0.001 | 0.505 | |
| Primary Substance (initial diagnosis) (%) | Cocaine hydrochloride | 3162 (3.7) | 957 (4.2) | 2205 (3.5) | |||
| Primary Substance (initial diagnosis) (%) | Cocaine paste | 3864 (4.5) | 1321 (5.9) | 2543 (4.1) | |||
| Primary Substance (initial diagnosis) (%) | Marijuana | 22991 (27.0) | 3083 (13.7) | 19908 (31.9) | |||
| Primary Substance (initial diagnosis) (%) | Other | 1987 (2.3) | 563 (2.5) | 1424 (2.3) | |||
| Primary Substance (initial diagnosis) (%) | [Missing] | 6421 (7.5) | 1054 (4.7) | 5367 (8.6) | |||
| Corrected birth year (median [IQR]) | 1980.00 [1971.00, 1987.00] | 1974.00 [1964.00, 1983.00] | 1982.00 [1974.00, 1988.00] | <0.001 | nonnorm | 0.630 | |
| Cohabitation status (Recoded) (f) (%) | Alone | 8026 (9.4) | 2765 (12.3) | 5261 (8.4) | <0.001 | 0.342 | |
| Cohabitation status (Recoded) (f) (%) | Family of origin | 35576 (41.8) | 6866 (30.4) | 28710 (45.9) | |||
| Cohabitation status (Recoded) (f) (%) | Others | 7291 (8.6) | 1802 (8.0) | 5489 (8.8) | |||
| Cohabitation status (Recoded) (f) (%) | With couple/children | 34154 (40.2) | 11118 (49.3) | 23036 (36.9) | |||
| Cohabitation status (Recoded) (f) (%) | [Missing] | 1 (0.0) | 1 (0.0) | 0 (0.0) | |||
| Physical Comorbidity (ICD-10) (%) | Without physical comorbidity | 33609 (39.5) | 8363 (37.1) | 25246 (40.4) | <0.001 | 0.084 | |
| Physical Comorbidity (ICD-10) (%) | Diagnosis unknown (under study) | 45892 (54.0) | 12457 (55.2) | 33435 (53.5) | |||
| Physical Comorbidity (ICD-10) (%) | One or more | 5547 (6.5) | 1732 (7.7) | 3815 (6.1) | |||
| Biopsychosocial compromise (%) | 1-Mild | 7986 (9.4) | 3801 (16.9) | 4185 (6.7) | <0.001 | 0.420 | |
| Biopsychosocial compromise (%) | 2-Moderate | 48355 (56.9) | 13544 (60.1) | 34811 (55.7) | |||
| Biopsychosocial compromise (%) | 3-Severe | 27150 (31.9) | 4743 (21.0) | 22407 (35.9) | |||
| Biopsychosocial compromise (%) | [Missing] | 1557 (1.8) | 464 (2.1) | 1093 (1.7) | |||
| Treatment Admission Motive (%) | Spontaneous | 38028 (44.7) | 10097 (44.8) | 27931 (44.7) | <0.001 | 0.187 | |
| Treatment Admission Motive (%) | Assisted Referral | 7967 (9.4) | 1455 (6.5) | 6512 (10.4) | |||
| Treatment Admission Motive (%) | Other | 4514 (5.3) | 1184 (5.3) | 3330 (5.3) | |||
| Treatment Admission Motive (%) | Justice Sector | 7976 (9.4) | 2787 (12.4) | 5189 (8.3) | |||
| Treatment Admission Motive (%) | Health Sector | 26563 (31.2) | 7029 (31.2) | 19534 (31.3) | |||
| tipo_centro_pub (%) | FALSE | 24070 (28.3) | 3414 (15.1) | 20656 (33.1) | <0.001 | 0.429 | |
| tipo_centro_pub (%) | TRUE | 60958 (71.7) | 19135 (84.8) | 41823 (66.9) | |||
| tipo_centro_pub (%) | [Missing] | 20 (0.0) | 3 (0.0) | 17 (0.0) |
Code
#kable(size=10, format="html",caption= "Summary descriptives, by Baseline Treatment Status") %>% kableExtra::kable_classic()
as.data.frame.TableOne(tbone_desc_merge_total_grant_23_24, smd=T, nonnormal= T)%>%
dplyr::mutate(char2=characteristic) %>%
tidyr::fill(char2) %>%
dplyr::select(char2,everything()) %>%
dplyr::mutate(level=ifelse(is.na(level),"[Missing]",level)) %>%
dplyr::mutate(char2=dplyr::case_when(characteristic=="NA"~NA_character_,TRUE~as.character(characteristic))) %>%
format_cells(1, 1:length(names(.)), "bold") %>%
dplyr::select(-1) %>%
dplyr::mutate(across(.fns = ~ str_replace_all(., pattern = "\\( ", replacement = "\\("))) %>%
dplyr::mutate(across(.fns = ~ str_trim(.))) %>%
write.table("_proposal_grant/2023/baseline_psu_desc_smd.csv", dec=",", sep="\t")Code
#C:\Users\CISS Fondecyt\Mi unidad\Alvacast\SISTRAT 2022 (github)\_proposal_grant/2023Long format
Code
CONS_C1_df_dup_SEP_2020_mod<-
CONS_C1_df_dup_SEP_2020[,c("hash_key","rn_hash","otras_sus1_mod","otras_sus2_mod","otras_sus3_mod","sus_principal_mod")] %>%
dplyr::mutate(across(starts_with("otras_sus"), ~ dplyr::case_when(
as.character(.) == "Cocaína" ~ "Cocaine hydrochloride",
as.character(.) == "Marihuana" ~ "Marijuana",
as.character(.) == "Otros" ~ "Other",
as.character(.) == "Pasta Base" ~ "Cocaine paste",
TRUE ~ as.character(.)
))) %>%
dplyr::mutate(sus_principal_mod= as.character(sus_principal_mod)) %>%
# Replace with missing if there are any duplicates
dplyr::mutate(otras_sus3_mod= dplyr::case_when(otras_sus3_mod==otras_sus2_mod| otras_sus3_mod==otras_sus1_mod| otras_sus3_mod==sus_principal_mod~ NA_character_, TRUE~ otras_sus3_mod)) %>%
dplyr::mutate(otras_sus2_mod= dplyr::case_when(otras_sus2_mod==otras_sus1_mod| otras_sus2_mod== sus_principal_mod~ NA_character_, TRUE~ otras_sus2_mod)) %>%
dplyr::mutate(otras_sus1_mod= dplyr::case_when(otras_sus1_mod== sus_principal_mod~ NA_character_, TRUE~ otras_sus1_mod)) %>%
# Empty substances were replaced with the previous
dplyr::mutate(otras_sus2_mod= dplyr::case_when(!is.na(otras_sus3_mod) & is.na(otras_sus2_mod)~ otras_sus3_mod, TRUE~ otras_sus2_mod)) %>%
dplyr::mutate(otras_sus1_mod= dplyr::case_when(!is.na(otras_sus2_mod) & is.na(otras_sus1_mod)~ otras_sus2_mod, TRUE~ otras_sus1_mod)) %>%
dplyr::mutate(sus_principal_mod= dplyr::case_when(!is.na(otras_sus1_mod) & is.na(sus_principal_mod)~ otras_sus1_mod, TRUE~ sus_principal_mod)) %>%
# Replace with missing if there are any duplicates
dplyr::mutate(otras_sus3_mod= dplyr::case_when(otras_sus3_mod==otras_sus2_mod| otras_sus3_mod==otras_sus1_mod| otras_sus3_mod==sus_principal_mod~ NA_character_, TRUE~ otras_sus3_mod)) %>%
dplyr::mutate(otras_sus2_mod= dplyr::case_when(otras_sus2_mod==otras_sus1_mod| otras_sus2_mod== sus_principal_mod~ NA_character_, TRUE~ otras_sus2_mod)) %>%
dplyr::mutate(otras_sus1_mod= dplyr::case_when(otras_sus1_mod== sus_principal_mod~ NA_character_, TRUE~ otras_sus1_mod)) %>%
# Replace with missing if there are any duplicates
dplyr::mutate(otras_sus3_mod= dplyr::case_when(otras_sus3_mod==otras_sus2_mod| otras_sus3_mod==otras_sus1_mod| otras_sus3_mod==sus_principal_mod~ NA_character_, TRUE~ otras_sus3_mod)) %>%
dplyr::mutate(otras_sus2_mod= dplyr::case_when(otras_sus2_mod==otras_sus1_mod| otras_sus2_mod== sus_principal_mod~ NA_character_, TRUE~ otras_sus2_mod)) %>%
dplyr::mutate(otras_sus1_mod= dplyr::case_when(otras_sus1_mod== sus_principal_mod~ NA_character_, TRUE~ otras_sus1_mod)) %>%
dplyr::select(hash_key, rn_hash, sus_principal_mod,otras_sus1_mod) %>%
dplyr::mutate(policonsumo2= ifelse(!is.na(otras_sus1_mod),1,0)) %>%
# dplyr::mutate(adv=ifelse(otras_sus1_mod==sus_principal_mod,1,0)) %>%
# dplyr::filter(adv==1)
data.table::data.table()Code
# vars_cov <- c("motivodeegreso_mod_imp_rec", "edad_al_ing_1", "sex", "edad_ini_cons", "escolaridad_rec", "sus_principal_mod", "freq_cons_sus_prin", "condicion_ocupacional_corr", "policonsumo", "otras_sus1_mod", "otras_sus2_mod", "otras_sus3_mod", "num_hijos_mod_joel_bin", "tenencia_de_la_vivienda_mod", "macrozona", "n_off_vio", "n_off_acq", "n_off_sud", "n_off_oth", "dg_cie_10_rec", "dg_trs_cons_sus_or", "clas_r", "porc_pobr", "sus_ini_mod_mvv", "ano_nac_corr", "con_quien_vive_joel", "fis_comorbidity_icd_10", "time_to_off_from_adm", "time_to_drop_from_adm", "age_at_censor_date", "age_tr_comp_imp", "age_offending_imp", "age_at_death","event_death", "n_post_off_vio", "n_post_off_acq", "n_post_off_sud", "n_post_off_oth", "n_post_off","hurto", "robo", "venta_drogas", "rina", "total_vif", "otro", "tot_off_top", "edad_a_ap_top_num", "edad_b_ap_top_num", "na_top_count", "n_top_records")
#tipo_de_plan_2_1
#motivodeegreso_mod_imp_1
#dias_treat_imp_sin_na_1
#fech_ing_num_1
#fech_egres_num_1
#edad_al_ing_1
#edad_al_egres_1
Base_fiscalia_v15f_grant_23_24_long<-
Base_fiscalia_v15f_grant_23_24 %>%
dplyr::select("hash_key",
paste0("motivodeegreso_mod_imp_",1:10),
paste0("dias_treat_imp_sin_na_",1:10),
paste0("fech_ing_num_",1:10),
paste0("fech_egres_num_",1:10),
paste0("tipo_de_plan_2_",1:10),
c("edad_al_ing_1", "sex", "edad_ini_cons", "escolaridad_rec", "sus_principal_mod", "freq_cons_sus_prin_ord", "condicion_ocupacional_corr24", "policonsumo", "num_hijos_mod_joel_bin", "tenencia_de_la_vivienda_mod", "macrozona", "n_off_vio", "n_off_acq", "n_off_sud", "n_off_oth", "dg_cie_10_rec", "dg_trs_cons_sus_or", "clas_r", "porc_pobr", "sus_ini_mod_mvv", "ano_nac_corr", "con_quien_vive_joel", "tipo_de_plan_2_mod",
#2024-01-28: se agregan 3 variables
"tipo_centro_pub", "origen_ingreso_mod", "compromiso_biopsicosocial")) %>%
#has 75 missing
dplyr::mutate(across(paste0("tipo_de_plan_2_",1:10),~dplyr::case_when(grepl("PR",.)~"residential",grepl("PAB",.)~"basic ambulatory",grepl("PAI",.)~"intensive ambulatory", T~NA_character_))) %>%
# Pivot longer
pivot_longer(
# 2024-01-27: por lo que entiendo, estas variables se excluyen de la transformación longitudinal. ES NECESARIO LONGITUDINALIZARLAS
cols = -c(hash_key, edad_al_ing_1, sex, edad_ini_cons, escolaridad_rec, sus_principal_mod, freq_cons_sus_prin_ord, condicion_ocupacional_corr24, policonsumo, num_hijos_mod_joel_bin, tenencia_de_la_vivienda_mod, macrozona, dg_cie_10_rec, dg_trs_cons_sus_or, clas_r, porc_pobr, sus_ini_mod_mvv, ano_nac_corr, con_quien_vive_joel, tipo_de_plan_2_mod,
#2024-01-28: se agregan 3 variables
tipo_centro_pub, origen_ingreso_mod, compromiso_biopsicosocial
),
names_to = c(".value", "treatment"),
names_pattern = "(.+)_(\\d+)$"
) %>%
dplyr::mutate(treatment=as.numeric(treatment)) %>%
dplyr::filter(!is.na(fech_ing_num)) %>%
# create event variables
arrange(hash_key, treatment) %>%
dplyr::group_by(hash_key) %>%
dplyr::mutate(enter_time=min(fech_ing_num, na.rm=T)) %>%
dplyr::mutate(time=fech_egres_num-enter_time) %>%
dplyr::mutate(cens_time= as.numeric(as.Date("2019-11-13"))-enter_time) %>%
dplyr::ungroup() %>%
#define event: dropout
dplyr::mutate(event=ifelse(motivodeegreso_mod_imp %in% c("Abandono Tardio", "Abandono Temprano", "Alta Administrativa"),1,0)) %>%
#censorship time if event not present
dplyr::mutate(time= dplyr::case_when(event==1~ cens_time, TRUE~ time)) %>% #to months
dplyr::mutate(time=time/30.5, cens_time= cens_time/30.5) %>%
#id variable
group_by(hash_key) %>%
mutate(is_first_occurrence = if_else(row_number() == 1, 1, 0)) %>%
ungroup() %>%
mutate(id = cumsum(is_first_occurrence)) %>%
dplyr::select(-is_first_occurrence)Code
# length(unique(Base_fiscalia_v15f_grant_23_24_long2$hash_key))
# [1] 72404
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#Added the time-varyinng polyuse
Base_fiscalia_v15f_grant_23_24_long2<-
Base_fiscalia_v15f_grant_23_24_long %>%
dplyr::left_join(CONS_C1_df_dup_SEP_2020_mod[,c("hash_key",
#2024-01-28: agregué biopsychosocial_comp
"rn_hash","policonsumo2")], by=c(c("hash_key"="hash_key","treatment"="rn_hash"))) %>% # nrow() 109756
dplyr::left_join(CONS_C1_df_dup_SEP_2020[,c("hash_key",
#2024-01-28: agregué biopsychosocial_comp
"rn_hash","compromiso_biopsicosocial")], by=c(c("hash_key"="hash_key","treatment"="rn_hash"))) %>% # nrow() 109756
dplyr::rename("compromiso_biopsicosocial"="compromiso_biopsicosocial.x") %>%
#2024-01-29: filtered treatments without finish
dplyr::filter(!motivodeegreso_mod_imp %in% c("En curso", "Muerte", "Derivación")) %>%
#dplyr::select(hash_key,fech_ing_num, fech_egres_num, time, enter_time, cens_time)
#2024-01-29: create event variables, again
arrange(hash_key, treatment) %>%
dplyr::group_by(hash_key) %>%
dplyr::mutate(treatment=row_number()) %>%
dplyr::mutate(max_treatment= max( treatment,na.rm=T)) %>%
dplyr::mutate(enter_time=min(fech_ing_num, na.rm=T)) %>%
dplyr::mutate(cens_time= as.numeric(as.Date("2019-11-13"))-enter_time) %>%
dplyr::ungroup() %>%
#2024-03-04: move out of the nested de discharge date
dplyr::mutate(time= fech_egres_num-enter_time) %>%
dplyr::mutate(time=time/30.5, cens_time= cens_time/30.5) %>%
dplyr::group_by(hash_key) %>%
dplyr::mutate(lag_time= lag(time)) %>%
dplyr::ungroup() %>%
dplyr::mutate(time=dplyr::case_when(lag_time==time & treatment==2~time+.001,
lag_time==time & treatment==3~time+.001,
lag_time==time & treatment==4~time+.001,
lag_time==time & treatment==5~time+.001,
lag_time==time & treatment==6~time+.001,
lag_time==time & treatment==7~time+.001,
lag_time==time & treatment==8~time+.001,
lag_time==time & treatment==9~time+.001,
TRUE~time)) %>%
dplyr::mutate(time=dplyr::case_when(lag_time==time & treatment==2~time+.001,
lag_time==time & treatment==3~time+.001,
lag_time==time & treatment==4~time+.001,
lag_time==time & treatment==5~time+.001,
lag_time==time & treatment==6~time+.001,
lag_time==time & treatment==7~time+.001,
lag_time==time & treatment==8~time+.001,
lag_time==time & treatment==9~time+.001,
TRUE~time)) %>%
dplyr::mutate(time=dplyr::case_when(lag_time==time & treatment==2~time+.001,
lag_time==time & treatment==3~time+.001,
lag_time==time & treatment==4~time+.001,
lag_time==time & treatment==5~time+.001,
lag_time==time & treatment==6~time+.001,
lag_time==time & treatment==7~time+.001,
lag_time==time & treatment==8~time+.001,
lag_time==time & treatment==9~time+.001,
TRUE~time)) %>%
dplyr::mutate(time=dplyr::case_when(lag_time==time & treatment==2~time+.001,
lag_time==time & treatment==3~time+.001,
lag_time==time & treatment==4~time+.001,
lag_time==time & treatment==5~time+.001,
lag_time==time & treatment==6~time+.001,
lag_time==time & treatment==7~time+.001,
lag_time==time & treatment==8~time+.001,
lag_time==time & treatment==9~time+.001,
TRUE~time)) %>%
dplyr::mutate(time=dplyr::case_when(lag_time==time & treatment==2~time+.001,
lag_time==time & treatment==3~time+.001,
lag_time==time & treatment==4~time+.001,
lag_time==time & treatment==5~time+.001,
lag_time==time & treatment==6~time+.001,
lag_time==time & treatment==7~time+.001,
lag_time==time & treatment==8~time+.001,
lag_time==time & treatment==9~time+.001,
TRUE~time)) %>%
dplyr::mutate(time=dplyr::case_when(lag_time==time & treatment==2~time+.001,
lag_time==time & treatment==3~time+.001,
lag_time==time & treatment==4~time+.001,
lag_time==time & treatment==5~time+.001,
lag_time==time & treatment==6~time+.001,
lag_time==time & treatment==7~time+.001,
lag_time==time & treatment==8~time+.001,
lag_time==time & treatment==9~time+.001,
TRUE~time)) %>%
dplyr::mutate(time=dplyr::case_when(lag_time==time & treatment==2~time+.001,
lag_time==time & treatment==3~time+.001,
lag_time==time & treatment==4~time+.001,
lag_time==time & treatment==5~time+.001,
lag_time==time & treatment==6~time+.001,
lag_time==time & treatment==7~time+.001,
lag_time==time & treatment==8~time+.001,
lag_time==time & treatment==9~time+.001,
TRUE~time)) %>%
dplyr::mutate(time=dplyr::case_when(lag_time==time & treatment==2~time+.001,
lag_time==time & treatment==3~time+.001,
lag_time==time & treatment==4~time+.001,
lag_time==time & treatment==5~time+.001,
lag_time==time & treatment==6~time+.001,
lag_time==time & treatment==7~time+.001,
lag_time==time & treatment==8~time+.001,
lag_time==time & treatment==9~time+.001,
TRUE~time)) %>%
#2024-01-29: censorship time if event not present
#2024-03-04: surv_time in other variable
dplyr::mutate(surv_time= dplyr::case_when(event==0~ cens_time, TRUE~ time))#to months
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_# TEST OVERLAPPINGS IN TIMES OF COUNTING PROCESS #_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
error_in_time_pre_followup<-
Base_fiscalia_v15f_grant_23_24_long2 %>%
dplyr::filter(dplyr::case_when(time <=lag_time & !is.na(lag_time)~T,T~F))
# A tibble: 3,031 x 36
Base_fiscalia_v15f_grant_23_24_long2 %>%
dplyr::filter(hash_key %in% unique(error_in_time_pre_followup$hash_key)) %>%
dplyr::select(hash_key, time, lag_time, motivodeegreso_mod_imp, dias_treat_imp_sin_na, enter_time, fech_ing_num, fech_egres_num, cens_time, treatment)# A tibble: 0 x 10
# ... with 10 variables: hash_key <chr>, time <dbl>, lag_time <dbl>,
# motivodeegreso_mod_imp <chr>, dias_treat_imp_sin_na <dbl>,
# enter_time <dbl>, fech_ing_num <dbl>, fech_egres_num <dbl>,
# cens_time <dbl>, treatment <int>Code
# A tibble: 7,706 x 8
# 0039f0d864e402fc7c9b8d326b6a
# 354 808 1296Code
tbone_desc_merge_total_grant_23_24_post<-
CreateTableOne(vars=c(setdiff(vars_cov, "policonsumo"), "offense_after_adm"), data= Base_fiscalia_v15f_grant_23_24[,c("hash_key", "motivodeegreso_mod_imp_rec", "edad_al_ing_1", "sex", "edad_ini_cons",
"escolaridad_rec", "sus_principal_mod", "freq_cons_sus_prin_ord",
"condicion_ocupacional_corr24", "policonsumo", "num_hijos_mod_joel_bin",
"tenencia_de_la_vivienda_mod", "macrozona", "dg_cie_10_rec", "dg_trs_cons_sus_or",
"clas_r", "porc_pobr", "sus_ini_mod_mvv", "ano_nac_corr", "con_quien_vive_joel",
"fis_comorbidity_icd_10", "compromiso_biopsicosocial", "origen_ingreso_mod","tipo_centro_pub")] %>% dplyr::filter(hash_key %in% unlist(unique(Base_fiscalia_v15f_grant_23_24_long2$hash_key))), factorVars = setdiff(vars_cov, c("motivodeegreso_mod_imp_rec","edad_al_ing_1", "edad_ini_cons","ano_nac_corr", "porc_pobr")), smd=T, strata="policonsumo", addOverall = T, includeNA=T, test=T)Warning in ModuleReturnVarsExist(vars, data): The data frame does not have: tipo_de_plan_2 tipo_de_plan_2_mod offense_after_adm Dropped
Warning in ModuleReturnVarsExist(factorVars, data): The data frame does not have: tipo_de_plan_2 tipo_de_plan_2_mod Dropped
Code
as.data.frame.TableOne(tbone_desc_merge_total_grant_23_24_post, smd=T, nonnormal= T)%>%
dplyr::mutate(char2=characteristic) %>%
tidyr::fill(char2) %>%
dplyr::select(char2,everything()) %>%
dplyr::mutate(level=ifelse(is.na(level),"[Missing]",level)) %>%
dplyr::mutate(char2=dplyr::case_when(characteristic=="NA"~NA_character_,TRUE~as.character(characteristic))) %>%
format_cells(1, 1:length(names(.)), "bold") %>%
dplyr::select(-1) %>%
dplyr::mutate(across(.fns = ~ str_replace_all(., pattern = "\\( ", replacement = "\\("))) %>%
dplyr::mutate(across(.fns = ~ str_trim(.))) %>%
knitr::kable(size=10, format="markdown",caption= "Summary descriptives, Polysubstance(1) and no Polysubstance use (0) [exclude ongoing treatments and external referrals]", escape=T)| characteristic | level | Overall | 0 | 1 | p | test | SMD |
|---|---|---|---|---|---|---|---|
| n | 72404 | 18707 | 53697 | ||||
| Complete status of treatment (binary) (%) | Treatment completion | 19278 (26.6) | 6404 (34.2) | 12874 (24.0) | <0.001 | 0.228 | |
| Complete status of treatment (binary) (%) | Treatment non-completion | 53117 (73.4) | 12303 (65.8) | 40814 (76.0) | |||
| Complete status of treatment (binary) (%) | [Missing] | 9 (0.0) | 0 (0.0) | 9 (0.0) | |||
| Age (admission to treatment) (median [IQR]) | 34.04 [27.38, 42.84] | 41.40 [32.14, 50.94] | 32.27 [26.47, 39.76] | <0.001 | nonnorm | 0.736 | |
| Sex (%) | Men | 55062 (76.0) | 13914 (74.4) | 41148 (76.6) | <0.001 | 0.052 | |
| Sex (%) | Women | 17342 (24.0) | 4793 (25.6) | 12549 (23.4) | |||
| Age of Onset of Substance Use (median [IQR]) | 15.00 [14.00, 18.00] | 17.00 [14.00, 20.00] | 15.00 [13.00, 17.00] | <0.001 | nonnorm | 0.419 | |
| Educational Attainment (%) | 1-More than high school | 11489 (15.9) | 2495 (13.3) | 8994 (16.7) | <0.001 | 0.236 | |
| Educational Attainment (%) | 2-Completed high school or less | 39904 (55.1) | 9289 (49.7) | 30615 (57.0) | |||
| Educational Attainment (%) | 3-Completed primary school or less | 20685 (28.6) | 6797 (36.3) | 13888 (25.9) | |||
| Educational Attainment (%) | [Missing] | 326 (0.5) | 126 (0.7) | 200 (0.4) | |||
| Primary Substance (admission to treatment) (%) | Alcohol | 24244 (33.5) | 11695 (62.5) | 12549 (23.4) | <0.001 | 0.877 | |
| Primary Substance (admission to treatment) (%) | Cocaine hydrochloride | 13505 (18.7) | 1906 (10.2) | 11599 (21.6) | |||
| Primary Substance (admission to treatment) (%) | Cocaine paste | 28563 (39.4) | 3922 (21.0) | 24641 (45.9) | |||
| Primary Substance (admission to treatment) (%) | Marijuana | 4835 (6.7) | 784 (4.2) | 4051 (7.5) | |||
| Primary Substance (admission to treatment) (%) | Other | 1256 (1.7) | 399 (2.1) | 857 (1.6) | |||
| Primary Substance (admission to treatment) (%) | [Missing] | 1 (0.0) | 1 (0.0) | 0 (0.0) | |||
| Frequency of Substance Use (Primary Substance) (%) | 1. Less than 1 day a week | 3554 (4.9) | 1276 (6.8) | 2278 (4.2) | <0.001 | 0.189 | |
| Frequency of Substance Use (Primary Substance) (%) | 2. 1 day a week or more | 4852 (6.7) | 1621 (8.7) | 3231 (6.0) | |||
| Frequency of Substance Use (Primary Substance) (%) | 3. 2 to 3 days a week | 20417 (28.2) | 5578 (29.8) | 14839 (27.6) | |||
| Frequency of Substance Use (Primary Substance) (%) | 4. 4 to 6 days a week | 11855 (16.4) | 2969 (15.9) | 8886 (16.5) | |||
| Frequency of Substance Use (Primary Substance) (%) | 5. Daily | 31363 (43.3) | 7145 (38.2) | 24218 (45.1) | |||
| Frequency of Substance Use (Primary Substance) (%) | [Missing] | 363 (0.5) | 118 (0.6) | 245 (0.5) | |||
| Corrected Occupational Status (%) | Employed | 35924 (49.6) | 10709 (57.2) | 25215 (47.0) | <0.001 | 0.275 | |
| Corrected Occupational Status (%) | Inactive | 11760 (16.2) | 3350 (17.9) | 8410 (15.7) | |||
| Corrected Occupational Status (%) | Unemployed | 24719 (34.1) | 4647 (24.8) | 20072 (37.4) | |||
| Corrected Occupational Status (%) | [Missing] | 1 (0.0) | 1 (0.0) | 0 (0.0) | |||
| Number of Children (dichotomized) (%) | 0 | 16802 (23.2) | 3356 (17.9) | 13446 (25.0) | <0.001 | 0.174 | |
| Number of Children (dichotomized) (%) | 1 | 54984 (75.9) | 15188 (81.2) | 39796 (74.1) | |||
| Number of Children (dichotomized) (%) | [Missing] | 618 (0.9) | 163 (0.9) | 455 (0.8) | |||
| tenencia_de_la_vivienda_mod (%) | Illegal Settlement | 763 (1.1) | 167 (0.9) | 596 (1.1) | <0.001 | 0.269 | |
| tenencia_de_la_vivienda_mod (%) | Others | 2041 (2.8) | 522 (2.8) | 1519 (2.8) | |||
| tenencia_de_la_vivienda_mod (%) | Owner/Transferred dwellings/Pays Dividends | 25339 (35.0) | 8171 (43.7) | 17168 (32.0) | |||
| tenencia_de_la_vivienda_mod (%) | Renting | 12354 (17.1) | 3145 (16.8) | 9209 (17.1) | |||
| tenencia_de_la_vivienda_mod (%) | Stays temporarily with a relative | 27759 (38.3) | 5637 (30.1) | 22122 (41.2) | |||
| tenencia_de_la_vivienda_mod (%) | [Missing] | 4148 (5.7) | 1065 (5.7) | 3083 (5.7) | |||
| Macro Administrative Zone in Chile (%) | Center | 54864 (75.8) | 13988 (74.8) | 40876 (76.1) | <0.001 | 0.292 | |
| Macro Administrative Zone in Chile (%) | North | 10690 (14.8) | 1882 (10.1) | 8808 (16.4) | |||
| Macro Administrative Zone in Chile (%) | South | 6833 (9.4) | 2835 (15.2) | 3998 (7.4) | |||
| Macro Administrative Zone in Chile (%) | [Missing] | 17 (0.0) | 2 (0.0) | 15 (0.0) | |||
| Psychiatric Comorbidity (ICD-10) (%) | Without psychiatric comorbidity | 28308 (39.1) | 8542 (45.7) | 19766 (36.8) | <0.001 | 0.187 | |
| Psychiatric Comorbidity (ICD-10) (%) | Diagnosis unknown (under study) | 13616 (18.8) | 2918 (15.6) | 10698 (19.9) | |||
| Psychiatric Comorbidity (ICD-10) (%) | With psychiatric comorbidity | 30480 (42.1) | 7247 (38.7) | 23233 (43.3) | |||
| SUD Severity (Dependence status) (%) | Drug dependence | 52416 (72.4) | 11498 (61.5) | 40918 (76.2) | <0.001 | 0.322 | |
| SUD Severity (Dependence status) (%) | Hazardous consumption | 19987 (27.6) | 7208 (38.5) | 12779 (23.8) | |||
| SUD Severity (Dependence status) (%) | [Missing] | 1 (0.0) | 1 (0.0) | 0 (0.0) | |||
| clas_r (%) | Mixed | 6938 (9.6) | 2359 (12.6) | 4579 (8.5) | <0.001 | 0.267 | |
| clas_r (%) | Rural | 5861 (8.1) | 2382 (12.7) | 3479 (6.5) | |||
| clas_r (%) | Urban | 59603 (82.3) | 13965 (74.7) | 45638 (85.0) | |||
| clas_r (%) | [Missing] | 2 (0.0) | 1 (0.0) | 1 (0.0) | |||
| Percentage of people in poverty (median [IQR]) | 0.12 [0.07, 0.17] | 0.11 [0.07, 0.16] | 0.12 [0.08, 0.17] | <0.001 | nonnorm | 0.105 | |
| Primary Substance (initial diagnosis) (%) | Alcohol | 39227 (54.2) | 12844 (68.7) | 26383 (49.1) | <0.001 | 0.506 | |
| Primary Substance (initial diagnosis) (%) | Cocaine hydrochloride | 2681 (3.7) | 769 (4.1) | 1912 (3.6) | |||
| Primary Substance (initial diagnosis) (%) | Cocaine paste | 3401 (4.7) | 1149 (6.1) | 2252 (4.2) | |||
| Primary Substance (initial diagnosis) (%) | Marijuana | 19644 (27.1) | 2589 (13.8) | 17055 (31.8) | |||
| Primary Substance (initial diagnosis) (%) | Other | 1648 (2.3) | 457 (2.4) | 1191 (2.2) | |||
| Primary Substance (initial diagnosis) (%) | [Missing] | 5803 (8.0) | 899 (4.8) | 4904 (9.1) | |||
| Corrected birth year (median [IQR]) | 1980.00 [1971.00, 1987.00] | 1973.00 [1964.00, 1983.00] | 1981.00 [1974.00, 1988.00] | <0.001 | nonnorm | 0.623 | |
| Cohabitation status (Recoded) (f) (%) | Alone | 6816 (9.4) | 2276 (12.2) | 4540 (8.5) | <0.001 | 0.337 | |
| Cohabitation status (Recoded) (f) (%) | Family of origin | 30074 (41.5) | 5648 (30.2) | 24426 (45.5) | |||
| Cohabitation status (Recoded) (f) (%) | Others | 6224 (8.6) | 1520 (8.1) | 4704 (8.8) | |||
| Cohabitation status (Recoded) (f) (%) | With couple/children | 29289 (40.5) | 9262 (49.5) | 20027 (37.3) | |||
| Cohabitation status (Recoded) (f) (%) | [Missing] | 1 (0.0) | 1 (0.0) | 0 (0.0) | |||
| Physical Comorbidity (ICD-10) (%) | Without physical comorbidity | 28661 (39.6) | 6946 (37.1) | 21715 (40.4) | <0.001 | 0.085 | |
| Physical Comorbidity (ICD-10) (%) | Diagnosis unknown (under study) | 39215 (54.2) | 10372 (55.4) | 28843 (53.7) | |||
| Physical Comorbidity (ICD-10) (%) | One or more | 4528 (6.3) | 1389 (7.4) | 3139 (5.8) | |||
| Biopsychosocial compromise (%) | 1-Mild | 6954 (9.6) | 3231 (17.3) | 3723 (6.9) | <0.001 | 0.418 | |
| Biopsychosocial compromise (%) | 2-Moderate | 41015 (56.6) | 11147 (59.6) | 29868 (55.6) | |||
| Biopsychosocial compromise (%) | 3-Severe | 22998 (31.8) | 3909 (20.9) | 19089 (35.5) | |||
| Biopsychosocial compromise (%) | [Missing] | 1437 (2.0) | 420 (2.2) | 1017 (1.9) | |||
| Treatment Admission Motive (%) | Spontaneous | 32847 (45.4) | 8462 (45.2) | 24385 (45.4) | <0.001 | 0.191 | |
| Treatment Admission Motive (%) | Assisted Referral | 6753 (9.3) | 1222 (6.5) | 5531 (10.3) | |||
| Treatment Admission Motive (%) | Other | 3935 (5.4) | 984 (5.3) | 2951 (5.5) | |||
| Treatment Admission Motive (%) | Justice Sector | 6813 (9.4) | 2380 (12.7) | 4433 (8.3) | |||
| Treatment Admission Motive (%) | Health Sector | 22056 (30.5) | 5659 (30.3) | 16397 (30.5) | |||
| tipo_centro_pub (%) | FALSE | 20677 (28.6) | 2806 (15.0) | 17871 (33.3) | <0.001 | 0.438 | |
| tipo_centro_pub (%) | TRUE | 51710 (71.4) | 15899 (85.0) | 35811 (66.7) | |||
| tipo_centro_pub (%) | [Missing] | 17 (0.0) | 2 (0.0) | 15 (0.0) |
Code
#kable(size=10, format="html",caption= "Summary descriptives, by Baseline Treatment Status") %>% kableExtra::kable_classic()
as.data.frame.TableOne(tbone_desc_merge_total_grant_23_24_post, smd=T, nonnormal= T)%>%
dplyr::mutate(char2=characteristic) %>%
tidyr::fill(char2) %>%
dplyr::select(char2,everything()) %>%
dplyr::mutate(level=ifelse(is.na(level),"[Missing]",level)) %>%
dplyr::mutate(char2=dplyr::case_when(characteristic=="NA"~NA_character_,TRUE~as.character(characteristic))) %>%
format_cells(1, 1:length(names(.)), "bold") %>%
dplyr::select(-1) %>%
dplyr::mutate(across(.fns = ~ str_replace_all(., pattern = "\\( ", replacement = "\\("))) %>%
dplyr::mutate(across(.fns = ~ str_trim(.))) %>%
write.table(paste0(getwd(),"/_proposal_grant/2023/baseline_psu_desc_smd_post.csv"), dec=",", sep="\t")Code
#C:\Users\CISS Fondecyt\Mi unidad\Alvacast\SISTRAT 2022 (github)\_proposal_grant/2023Code
set.seed(21251)
IrregLong::abacus.plot(
25,
"time",
"id",
Base_fiscalia_v15f_grant_23_24_long2%>% #dplyr::filter(event==1)%>%
dplyr::group_by(hash_key) %>%
dplyr::mutate(is_first_occurrence = if_else(row_number() == 1, 1, 0)) %>%
dplyr::ungroup() %>%
dplyr::mutate(id = cumsum(is_first_occurrence)) %>%
dplyr::select(-is_first_occurrence) %>% data.frame(),
0,
120,
xlab.abacus = "Time (in months of follow-up)",
ylab.abacus = "Subject",
pch.abacus = 16,
col.abacus = 1
)
Code
#abacus_plot_irrelong
jpeg(filename = "_proposal_grant/2023/_figs/abacus.jpg", units = "in", width=8, height=5, res=500)
set.seed(21251)
IrregLong::abacus.plot(
25,
"time",
"id",
Base_fiscalia_v15f_grant_23_24_long2%>% #dplyr::filter(event==1)%>%
dplyr::group_by(hash_key) %>%
dplyr::mutate(is_first_occurrence = if_else(row_number() == 1, 1, 0)) %>%
dplyr::ungroup() %>%
dplyr::mutate(id = cumsum(is_first_occurrence)) %>%
dplyr::select(-is_first_occurrence) %>% data.frame(),
0,
120,
xlab.abacus = "Time (in months of follow-up)",
ylab.abacus = "Subject",
pch.abacus = 16,
col.abacus = 1
)
dev.off()png
2 Code
summary(subset(Base_fiscalia_v15f_grant_23_24_long2, treatment==1)$max_treatment) Min. 1st Qu. Median Mean 3rd Qu. Max.
1.000 1.000 1.000 1.244 1.000 10.000 Code
#prop. q tienen sólo uno
cat("Only one treatment")Only one treatmentCode
scales::percent(table(subset(Base_fiscalia_v15f_grant_23_24_long2, treatment==1)$max_treatment)[1] / nrow(subset(Base_fiscalia_v15f_grant_23_24_long2, treatment==1)), accuracy=.1) 1
"81.6%" Code
#scales::percent(59087 /72404, accuracy=.1 )
cat("Patients with 4 or more treatments")Patients with 4 or more treatmentsCode
scales::percent(
sum(table(subset(Base_fiscalia_v15f_grant_23_24_long2, treatment==1)$max_treatment)[4:100], na.rm=T)/nrow(subset(Base_fiscalia_v15f_grant_23_24_long2, treatment==1))
)[1] "1%"Code
#[1] "1%"
Base_fiscalia_v15f_grant_23_24[,c("hash_key", "motivodeegreso_mod_imp_rec", "edad_al_ing_1", "sex", "edad_ini_cons",
"escolaridad_rec", "sus_principal_mod", "freq_cons_sus_prin_ord",
"condicion_ocupacional_corr24", "policonsumo", "num_hijos_mod_joel_bin",
"tenencia_de_la_vivienda_mod", "macrozona", "dg_cie_10_rec", "dg_trs_cons_sus_or",
"clas_r", "porc_pobr", "sus_ini_mod_mvv", "ano_nac_corr", "con_quien_vive_joel",
"fis_comorbidity_icd_10", "compromiso_biopsicosocial", "origen_ingreso_mod","tipo_centro_pub","duplicates_filtered")] %>% dplyr::filter(hash_key %in% unlist(unique(Base_fiscalia_v15f_grant_23_24_long2$hash_key))) %>% dplyr::mutate(duplicates_filtered= dplyr::case_when(as.numeric(duplicates_filtered)>3~4,TRUE~as.numeric(duplicates_filtered))) %>%
janitor::tabyl(policonsumo,duplicates_filtered)%>%
chisq.test()Warning in chisq.test(.): Chi-squared approximation may be incorrect
Pearson's Chi-squared test
data: .
X-squared = 692.64, df = 4, p-value < 2.2e-16Code
#X-squared = 693, df = 4, p-value <2e-16
data.frame(
col1 = c(15527, 39686),
col2 = c(2475, 9902),
col3 = c(530, 2841),
col4 = c(175, 1268)
)%>%
mutate(across(everything(), ~ . / sum(.)) * 100) %>%
mutate(across(everything(), round, digits = 2)) col1 col2 col3 col4
1 28.12 20 15.72 12.13
2 71.88 80 84.28 87.87Code
# col1 col2 col3 col4
# 1 28.12 20 15.72 12.13
# 2 71.88 80 84.28 87.87
invisible("Patients with more treatments have more baseline PSU")
cbind.data.frame(motivodeegreso_mod_imp_rec = c("Treatment completion",
"Treatment non-completion"),
`1` = c(15782, 39428), `2` = c(2622, 9750),
`3` = c(654, 2717), `4` = c(220, 1222)) %>%
mutate_if(is.numeric,~ . / sum(.) * 100) motivodeegreso_mod_imp_rec 1 2 3 4
1 Treatment completion 28.5854 21.19302 19.40077 15.25659
2 Treatment non-completion 71.4146 78.80698 80.59923 84.74341Code
# motivodeegreso_mod_imp_rec 1 2 3 4
# 1 Treatment completion 28.59 21.19 19.4 15.26
# 2 Treatment non-completion 71.41 78.81 80.6 84.74 As seen in the previous plot, a few observations had more than one treatment.
Code
paste0("Patients with more than one value of PSU (changes in patterns): ",
Base_fiscalia_v15f_grant_23_24_long2%>%
dplyr::group_by(hash_key) %>%
dplyr::group_by(id) %>%
summarise(n_dis_psu=n_distinct(policonsumo2), n_dis_event=n_distinct(event), n_dis_event=n_distinct(dias_treat_imp_sin_na), max_tr=max(treatment)) %>%
dplyr::filter(n_dis_psu>1) %>% nrow()
) [1] "Patients with more than one value of PSU (changes in patterns): 3235"Code
#[1] "Patients with more than one value of PSU (changes in patterns): 3235"
paste0("Patients with more than one value of event: ",
Base_fiscalia_v15f_grant_23_24_long2%>%
dplyr::group_by(id) %>%
summarise(n_dis_psu=n_distinct(policonsumo2), n_dis_event=n_distinct(event), n_dis_dias_tr=n_distinct(dias_treat_imp_sin_na), max_tr=max(treatment)) %>%
dplyr::filter(n_dis_event>1) %>% nrow()
)[1] "Patients with more than one value of event: 4623"Code
#[1] "Patients with more than one value of event: 4623"
paste0("Patients with more than one value of days in treatment: ",
Base_fiscalia_v15f_grant_23_24_long2%>%
dplyr::group_by(id) %>%
summarise(n_dis_psu=n_distinct(policonsumo2), n_dis_event=n_distinct(event), n_dis_dias_tr=n_distinct(dias_treat_imp_sin_na), max_tr=max(treatment)) %>%
dplyr::filter(n_dis_dias_tr>1) %>% nrow()
)[1] "Patients with more than one value of days in treatment: 13289"Code
#[1] "Patients with more than one value of days in treatment: 13289"
paste0("Patients with more than one value in biopsychosocial compromise: ",
Base_fiscalia_v15f_grant_23_24_long2%>%
dplyr::group_by(id) %>%
summarise(n_dis_psu=n_distinct(policonsumo2), n_dis_event=n_distinct(event), n_dis_dias_tr=n_distinct(dias_treat_imp_sin_na), n_dis_biopsicomp= n_distinct(compromiso_biopsicosocial.y), max_tr=max(treatment)) %>%
dplyr::filter(n_dis_biopsicomp>1) %>% nrow()
)[1] "Patients with more than one value in biopsychosocial compromise: 6691"Code
#[1] "Patients with more than one value in biopsychosocial compromise: 6691"Code
library(easyalluvial)
library(parcats)
#tipo_de_plan2
#treatment, dias_treat_imp_sin_na, policonsumo2, event
# basic ambulatory intensive ambulatory residential
table(subset(Base_fiscalia_v15f_grant_23_24_long2, treatment==1)$max_treatment)
1 2 3 4 5 6 7 8 10
59087 10121 2363 601 163 52 12 4 1 Code
cat("================================")================================Code
invisible("First 4 treatments, stratified by treatment plan, to see if polysubstance use changes by treatment")
# subset(Base_fiscalia_v15f_grant_23_24_long2, hash_key%in% (dplyr::filter(Base_fiscalia_v15f_grant_23_24_long2, tipo_de_plan_2=="residential" & treatment==1) %>% dplyr::pull("hash_key") %>% unique()) & treatment<5,c("id","treatment","motivodeegreso_mod_imp","policonsumo2"))
p_alluvial<-
alluvial_long(data.frame(subset(Base_fiscalia_v15f_grant_23_24_long2, treatment<5) %>% dplyr::mutate(policonsumo2=if_else(policonsumo2==1,"PSU","no PSU",NA_character_))),
key= treatment,
id= id,
value= policonsumo2,
fill_by="first_variable",
NA_label = "censored",
complete=F,
bin=3,
stratum_label_size=2,
stratum_width=1/4,
bin_labels = c("no PSU", "PSU"),
col_vector_flow=c("red","gray60"),
col_vector_value=c("#bf6f6f","#eec1ad","gray60") )+
theme_void()
#1: In .f(.x[[i]], ...) : bins (0.4544,0.4549] of policonsumo2 are empty
p_alluvial$data$x_numeric <- as.numeric(as.character(p_alluvial$data$x)) # Convert factor to numeric correctly
library(ggalluvial)Warning: package ‘ggalluvial’ was built under R version 4.1.3
Code
if( !is.na(match("compute",ls())) ){
p_alluvial$data %>%
# dplyr::group_by(x)%>%
# dplyr::mutate(Percentage = round(n / sum(n[value==]) * 100,1)) %>%
# dplyr::mutate(print=glue::glue("{format(n, big.mark=',')}; {Percentage}%"))%>%
dplyr::ungroup() %>%
ggplot(
aes(axis1 = x_numeric, axis2 = x_numeric + 1, y = n)) +#, label = print
geom_alluvium(aes(fill = fill)) +
geom_stratum() +
geom_text(stat = "stratum", size = 3) + # Adjust size as needed
theme_minimal() +
ggtitle("Alluvial Plot with Custom Flow Labels") +
xlab("Stage") +
ylab("Count")
}
df_alluvial<-
p_alluvial$data_key %>% group_by(alluvial_id) %>% slice(1) %>%
dplyr::group_by(`1`) %>%
dplyr::mutate(perc1=scales::percent(n/sum(n), accuracy=.1)) %>%
dplyr::group_by(`1`,`2`)%>%
dplyr::mutate(perc2=scales::percent(n/sum(n), accuracy=.1)) %>%
ungroup() %>%
data.table::data.table()
plot2<-
ggplot(data =df_alluvial,
aes(axis1 = `1`, axis2 = `2`, axis3 = `3`, y = n)) +
geom_alluvium(aes(fill = `1`)) +
geom_stratum(alpha = .5) +
geom_text(stat = "stratum", aes(label = paste0(after_stat(stratum),"; ",scales::percent(after_stat(prop), .1))), size=4) +
# geom_text(stat = "flow", nudge_x = 0.3,
# aes(label = paste0(after_stat(flow)," ",round(after_stat(count)*after_stat(prop), 0)),
# hjust = (after_stat(flow) == "to")),
# size=3) +
scale_fill_manual(name="PSU",values = c("no PSU" = "darkred", "PSU" = "darkblue", "censored" = "gray60")) +
xlab("Stage") +
theme_void()+
theme(legend.position="bottom")
ggsave(plot2, file="_proposal_grant/2023/_figs/alluvial.png", width=8.5, height=5.5, dpi=500)Warning in to_lodes_form(data = data, axes = axis_ind, discern = params$discern): Some strata appear at multiple axes.
Warning in to_lodes_form(data = data, axes = axis_ind, discern = params$discern): Some strata appear at multiple axes.
Warning in to_lodes_form(data = data, axes = axis_ind, discern = params$discern): Some strata appear at multiple axes.
Code
ggsave(plot2, file="_proposal_grant/2023/_figs/alluvial.jpg", width=8.5, height=5.5, dpi=500)Warning in to_lodes_form(data = data, axes = axis_ind, discern = params$discern): Some strata appear at multiple axes.
Warning in to_lodes_form(data = data, axes = axis_ind, discern = params$discern): Some strata appear at multiple axes.
Warning in to_lodes_form(data = data, axes = axis_ind, discern = params$discern): Some strata appear at multiple axes.
Code
newdat <- layer_data(plot2)Warning in to_lodes_form(data = data, axes = axis_ind, discern = params$discern): Some strata appear at multiple axes.
Warning in to_lodes_form(data = data, axes = axis_ind, discern = params$discern): Some strata appear at multiple axes.
Warning in to_lodes_form(data = data, axes = axis_ind, discern = params$discern): Some strata appear at multiple axes.
Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
cat("======================================")======================================Code
invisible("Database to summarise trajectories")
# data.frame(subset(Base_fiscalia_v15f_grant_23_24_long2, hash_key%in% (dplyr::filter(Base_fiscalia_v15f_grant_23_24_long2, tipo_de_plan_2=="basic ambulatory" & treatment==1) %>% dplyr::pull("hash_key") %>% unique()) & treatment<5,c("id","treatment","dias_treat_imp_sin_na","motivodeegreso_mod_imp","policonsumo2")) %>% dplyr::mutate(policonsumo2= if_else(policonsumo2==1,"PSU","no PSU",NA_character_))) %>%
df_long_trajectories_pre<-
data.frame(subset(Base_fiscalia_v15f_grant_23_24_long2, hash_key %in% (dplyr::filter(Base_fiscalia_v15f_grant_23_24_long2, treatment==1) %>% dplyr::pull("hash_key") %>% unique()) & treatment<5,c("id","treatment","dias_treat_imp_sin_na","motivodeegreso_mod_imp","policonsumo2")) %>% dplyr::mutate(policonsumo2= if_else(policonsumo2==1,"PSU","no PSU",NA_character_))) %>%
dplyr::mutate(motivodeegreso_mod_imp= dplyr::case_when(dias_treat_imp_sin_na<90 & grepl("Admin",motivodeegreso_mod_imp)~ "Abandono Temprano", dias_treat_imp_sin_na>=90 & grepl("Admin",motivodeegreso_mod_imp)~ "Abandono Tardio", T~ motivodeegreso_mod_imp)) %>%
dplyr::select(-dias_treat_imp_sin_na) %>%
dplyr::filter(!motivodeegreso_mod_imp %in% c("Derivación","En curso", "Muerte")) %>%
dplyr::filter(!is.na(motivodeegreso_mod_imp)) %>%
dplyr::ungroup()
cat("======================================")======================================Code
invisible("Database to summarise trajectories until 4th treatment")
df_long_trajectories7<-
df_long_trajectories_pre %>%
dplyr::group_by(id) %>%
dplyr::summarise(trajectory = paste(policonsumo2, motivodeegreso_mod_imp, sep = " -> ", collapse = " -> "), .groups = 'drop') %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= n(),.groups="drop") %>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
dplyr::filter(repeats==max(repeats)) %>%
dplyr::mutate(trajectory2= trajectory) %>%
tidyr::separate(trajectory2, sep=" -> ", into=paste0("traj",1:(max(.$repeats)+1))) %>%
tidyr::unite("trajectory2",paste0("traj",1:(max(.$repeats))), sep= " -> ", na.rm=T) %>%
dplyr::group_by(trajectory2) %>%
dplyr::arrange(desc(count))%>%
dplyr::mutate(Percentage = round(count / sum(count) * 100,1)) %>%
dplyr::mutate(print=glue::glue("{format(count, big.mark=',')}; {Percentage}%")) %>%
dplyr::ungroup()
df_long_trajectories6<-
df_long_trajectories_pre %>%
dplyr::group_by(id) %>%
dplyr::summarise(trajectory = paste(policonsumo2, motivodeegreso_mod_imp, sep = " -> ", collapse = " -> "), .groups = 'drop') %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= n(),.groups="drop") %>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
tidyr::separate(trajectory, sep=" -> ", into=paste0("traj",1:(max(.$repeats)+1))) %>%
tidyr::unite("trajectory",paste0("traj",1:(max(.$repeats))), sep= " -> ", na.rm=T) %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= sum(count, na.rm=T),.groups="drop") %>%
dplyr::arrange(desc(count))%>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
dplyr::filter(repeats==max(repeats)) %>%
dplyr::mutate(trajectory2= trajectory) %>%
tidyr::separate(trajectory2, sep=" -> ", into=paste0("traj",1:(max(.$repeats)+1))) %>%
tidyr::unite("trajectory2",paste0("traj",1:(max(.$repeats))), sep= " -> ", na.rm=T) %>%
dplyr::group_by(trajectory2) %>%
dplyr::arrange(desc(count))%>%
dplyr::mutate(Percentage = round(count / sum(count) * 100,1)) %>%
dplyr::mutate(print=glue::glue("{format(count, big.mark=',')}; {Percentage}%")) %>%
dplyr::ungroup()
df_long_trajectories5<-
df_long_trajectories_pre %>%
dplyr::filter(treatment <4) %>%
dplyr::group_by(id) %>%
dplyr::summarise(trajectory = paste(policonsumo2, motivodeegreso_mod_imp, sep = " -> ", collapse = " -> "), .groups = 'drop') %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= n(),.groups="drop") %>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
dplyr::filter(repeats==max(repeats)) %>%
dplyr::mutate(trajectory2= trajectory) %>%
tidyr::separate(trajectory2, sep=" -> ", into=paste0("traj",1:(max(.$repeats)+1))) %>%
tidyr::unite("trajectory2",paste0("traj",1:(max(.$repeats))), sep= " -> ", na.rm=T) %>%
dplyr::group_by(trajectory2) %>%
dplyr::arrange(desc(count))%>%
dplyr::mutate(Percentage = round(count / sum(count) * 100,1)) %>%
dplyr::mutate(print=glue::glue("{format(count, big.mark=',')}; {Percentage}%")) %>%
dplyr::ungroup()
df_long_trajectories4<-
df_long_trajectories_pre %>%
dplyr::filter(treatment <4) %>%
dplyr::group_by(id) %>%
dplyr::summarise(trajectory = paste(policonsumo2, motivodeegreso_mod_imp, sep = " -> ", collapse = " -> "), .groups = 'drop') %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= n(),.groups="drop") %>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
tidyr::separate(trajectory, sep=" -> ", into=paste0("traj",1:(max(.$repeats)+1))) %>%
tidyr::unite("trajectory",paste0("traj",1:(max(.$repeats))), sep= " -> ", na.rm=T) %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= sum(count, na.rm=T),.groups="drop") %>%
dplyr::arrange(desc(count))%>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
dplyr::filter(repeats==max(repeats)) %>%
dplyr::mutate(trajectory2= trajectory) %>%
tidyr::separate(trajectory2, sep=" -> ", into=paste0("traj",1:(max(.$repeats)+1))) %>%
tidyr::unite("trajectory2",paste0("traj",1:(max(.$repeats))), sep= " -> ", na.rm=T) %>%
dplyr::group_by(trajectory2) %>%
dplyr::arrange(desc(count))%>%
dplyr::mutate(Percentage = round(count / sum(count) * 100,1)) %>%
dplyr::mutate(print=glue::glue("{format(count, big.mark=',')}; {Percentage}%")) %>%
dplyr::ungroup()
df_long_trajectories3<-
df_long_trajectories_pre %>%
dplyr::filter(treatment <3) %>%
dplyr::group_by(id) %>%
dplyr::summarise(trajectory = paste(policonsumo2, motivodeegreso_mod_imp, sep = " -> ", collapse = " -> "), .groups = 'drop') %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= n(),.groups="drop") %>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
dplyr::filter(repeats==max(repeats)) %>%
dplyr::mutate(trajectory2= trajectory) %>%
tidyr::separate(trajectory2, sep=" -> ", into=paste0("traj",1:(max(.$repeats)+1))) %>%
tidyr::unite("trajectory2",paste0("traj",1:(max(.$repeats))), sep= " -> ", na.rm=T) %>%
dplyr::group_by(trajectory2) %>%
dplyr::arrange(desc(count))%>%
dplyr::mutate(Percentage = round(count / sum(count) * 100,1)) %>%
dplyr::mutate(print=glue::glue("{format(count, big.mark=',')}; {Percentage}%")) %>%
dplyr::ungroup()
df_long_trajectories2<-
df_long_trajectories_pre %>%
dplyr::filter(treatment <3) %>%
dplyr::group_by(id) %>%
dplyr::summarise(trajectory = paste(policonsumo2, motivodeegreso_mod_imp, sep = " -> ", collapse = " -> "), .groups = 'drop') %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= n(),.groups="drop") %>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
tidyr::separate(trajectory, sep=" -> ", into=paste0("traj",1:(max(.$repeats)+1))) %>%
tidyr::unite("trajectory",paste0("traj",1:(max(.$repeats))), sep= " -> ", na.rm=T) %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= sum(count, na.rm=T),.groups="drop") %>%
dplyr::arrange(desc(count))%>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
dplyr::filter(repeats==max(repeats)) %>%
dplyr::mutate(trajectory2= trajectory) %>%
tidyr::separate(trajectory2, sep=" -> ", into=paste0("traj",1:(max(.$repeats)+1))) %>%
tidyr::unite("trajectory2",paste0("traj",1:(max(.$repeats))), sep= " -> ", na.rm=T) %>%
dplyr::group_by(trajectory2) %>%
dplyr::arrange(desc(count))%>%
dplyr::mutate(Percentage = round(count / sum(count) * 100,1)) %>%
dplyr::mutate(print=glue::glue("{format(count, big.mark=',')}; {Percentage}%")) %>%
dplyr::ungroup()
df_long_trajectories1<-
df_long_trajectories_pre %>%
dplyr::filter(treatment <2) %>%
dplyr::group_by(id) %>%
dplyr::summarise(trajectory = paste(policonsumo2, motivodeegreso_mod_imp, sep = " -> ", collapse = " -> "), .groups = 'drop') %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= n(),.groups="drop") %>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
dplyr::filter(repeats==max(repeats)) %>%
dplyr::mutate(trajectory2= trajectory) %>%
tidyr::separate(trajectory2, sep=" -> ", into=paste0("traj",1:(max(.$repeats)+1))) %>%
tidyr::unite("trajectory2",paste0("traj",1:(max(.$repeats))), sep= " -> ", na.rm=T) %>%
dplyr::group_by(trajectory2) %>%
dplyr::arrange(desc(count))%>%
dplyr::mutate(Percentage = round(count / sum(count) * 100,1)) %>%
dplyr::mutate(print=glue::glue("{format(count, big.mark=',')}; {Percentage}%")) %>%
dplyr::ungroup()
df_long_trajectories0<-
df_long_trajectories_pre %>%
dplyr::filter(treatment <2) %>%
dplyr::group_by(id) %>%
dplyr::summarise(trajectory = paste(policonsumo2, motivodeegreso_mod_imp, sep = " -> ", collapse = " -> "), .groups = 'drop') %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= n(),.groups="drop") %>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
tidyr::separate(trajectory, sep=" -> ", into=paste0("traj",1:(max(.$repeats)+1))) %>%
tidyr::unite("trajectory",paste0("traj",1:(max(.$repeats))), sep= " -> ", na.rm=T) %>%
dplyr::group_by(trajectory) %>%
dplyr::summarise(count= sum(count, na.rm=T),.groups="drop") %>%
dplyr::arrange(desc(count))%>%
dplyr::mutate(repeats = str_count(trajectory, pattern = "->")) %>%
dplyr::filter(repeats==max(repeats)) %>%
dplyr::arrange(desc(count)) %>%
dplyr::mutate(Percentage = round(count / sum(count) * 100,1)) %>%
dplyr::mutate(print=glue::glue("{format(count, big.mark=',')}; {Percentage}%")) %>%
dplyr::ungroup()
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
cbind.fill <- function(...) {
transpoted <- lapply(list(...),t)
transpoted_dataframe <- lapply(transpoted, as.data.table, keep.rownames=T)
return (data.frame(t(plyr::rbind.fill(transpoted_dataframe))))
}
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
tree_str<-
plyr::rbind.fill(cbind(step=1,df_long_trajectories0),
cbind(step=2,df_long_trajectories1),
cbind(step=3,df_long_trajectories2),
cbind(step=4,df_long_trajectories3),
cbind(step=5,df_long_trajectories4),
cbind(step=6,df_long_trajectories5),
cbind(step=7,df_long_trajectories6),
cbind(step=8,df_long_trajectories7)) %>%
dplyr::mutate(trajectory2 = str_replace_all(trajectory, c(
"no PSU" = "0", #must be first because it is more specific
"PSU" = "1",
"Abandono Temprano" = "0",
"Abandono Tardio" = "1",
"Alta Terapéutica" = "2",
" -> " = "",
" " = ""
)))
source(paste0(getwd(),"/_proposal_grant/2023/tree_str_pab.R"))
#tree_str_pab
source(paste0(getwd(),"/_proposal_grant/2023/tree_str_pai.R"))
#tree_str_pai
source(paste0(getwd(),"/_proposal_grant/2023/tree_str_pr.R"))
#tree_str_prCode
library(data.tree)
traveller <- Node$new("Admission")
route0 <- traveller$AddChild(paste0("1st tr., no PSU:\n", as.character(subset(tree_str, trajectory2=="0", "print"))
))
route1 <- traveller$AddChild(paste0("1st tr., PSU:\n", as.character(subset(tree_str, trajectory2=="1", "print"))
))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#step 1
route00 <- route0$AddChild(paste0("1st tr., no PSU, Early drop:\n",
as.character(subset(tree_str, trajectory2=="00", "print"))
))
route01 <- route0$AddChild(paste0("1st tr., no PSU, Late drop:\n",
as.character(subset(tree_str, trajectory2=="01", "print"))
))
route02 <- route0$AddChild(paste0("1st tr., no PSU, Tr. comp:\n",
as.character(subset(tree_str, trajectory2=="02", "print"))
))
route10 <- route1$AddChild(paste0("1st tr., PSU, Early drop:\n",
as.character(subset(tree_str, trajectory2=="10", "print"))
))
route11 <- route1$AddChild(paste0("1st tr., PSU, Late drop:\n",
as.character(subset(tree_str, trajectory2=="11", "print"))
))
route12 <- route1$AddChild(paste0("1st tr., PSU, Tr. comp:\n",
as.character(subset(tree_str, trajectory2=="12", "print"))
))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#step 2
route000 <- route00$AddChild(paste0("2nd tr., no PSU:\n", as.character(subset(tree_str, trajectory2=="000", "print"))
))
route001 <- route00$AddChild(paste0("2nd tr., PSU:\n", as.character(subset(tree_str, trajectory2=="001", "print"))
))
route010 <- route01$AddChild(paste0("2nd tr., no PSU:\n", as.character(subset(tree_str, trajectory2=="010", "print"))
))
route011 <- route01$AddChild(paste0("2nd tr., PSU:\n", as.character(subset(tree_str, trajectory2=="011", "print"))
))
route020 <- route02$AddChild(paste0("2nd tr., no PSU:\n", as.character(subset(tree_str, trajectory2=="020", "print"))
))
route021 <- route02$AddChild(paste0("2nd tr., PSU:\n", as.character(subset(tree_str, trajectory2=="021", "print"))
))
route100 <- route10$AddChild(paste0("2nd tr., no PSU:\n", as.character(subset(tree_str, trajectory2=="100", "print"))
))
route101 <- route10$AddChild(paste0("2nd tr., PSU:\n", as.character(subset(tree_str, trajectory2=="101", "print"))
))
route110 <- route11$AddChild(paste0("2nd tr., no PSU:\n", as.character(subset(tree_str, trajectory2=="110", "print"))
))
route111 <- route11$AddChild(paste0("2nd tr., PSU:\n", as.character(subset(tree_str, trajectory2=="111", "print"))
))
route120 <- route12$AddChild(paste0("2nd tr., no PSU:\n", as.character(subset(tree_str, trajectory2=="120", "print"))
))
route121 <- route12$AddChild(paste0("2nd tr., PSU:\n", as.character(subset(tree_str, trajectory2=="121", "print"))
))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#step 3
route0000 <- route000$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="0000", "print"))
))
route0010 <- route001$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="0010", "print"))
))
route0100 <- route010$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="0100", "print"))
))
route0110 <- route011$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="0110", "print"))
))
route0200 <- route020$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="0200", "print"))
))
route0210 <- route021$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="0210", "print"))
))
route1000 <- route100$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="1000", "print"))
))
route1010 <- route101$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="1010", "print"))
))
route1100 <- route110$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="1100", "print"))
))
route1110 <- route111$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="1110", "print"))
))
route1200 <- route120$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="1200", "print"))
))
route1210 <- route121$AddChild(paste0("3rd tr., Early drop:\n", as.character(subset(tree_str, trajectory2=="1210", "print"))
))
route0001 <- route000$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="0001", "print"))
))
route0011 <- route001$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="0011", "print"))
))
route0101 <- route010$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="0101", "print"))
))
route0111 <- route011$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="0111", "print"))
))
route0201 <- route020$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="0201", "print"))
))
route0211 <- route021$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="0211", "print"))
))
route1001 <- route100$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="1001", "print"))
))
route1011 <- route101$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="1011", "print"))
))
route1101 <- route110$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="1101", "print"))
))
route1111 <- route111$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="1111", "print"))
))
route1201 <- route120$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="1201", "print"))
))
route1211 <- route121$AddChild(paste0("3rd tr., Late drop:\n", as.character(subset(tree_str, trajectory2=="1211", "print"))
))
route0002 <- route000$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="0002", "print"))
))
route0012 <- route001$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="0012", "print"))
))
route0102 <- route010$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="0102", "print"))
))
route0112 <- route011$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="0112", "print"))
))
route0202 <- route020$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="0202", "print"))
))
route0212 <- route021$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="0212", "print"))
))
route1002 <- route100$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="1002", "print"))
))
route1012 <- route101$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="1012", "print"))
))
route1102 <- route110$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="1102", "print"))
))
route1112 <- route111$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="1112", "print"))
))
route1202 <- route120$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="1202", "print"))
))
route1212 <- route121$AddChild(paste0("3rd tr., Tr. comp:\n", as.character(subset(tree_str, trajectory2=="1212", "print"))
))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
SetGraphStyle(traveller, rankdir = "LR")
plot(traveller)Missingness
Code
paste0("Percentage of the total that has missing values: ",
scales::percent(Base_fiscalia_v15f_grant_23_24_long2[which(complete.cases(subset(Base_fiscalia_v15f_grant_23_24_long2, select= c(hash_key, tipo_de_plan_2, tipo_de_plan_2_mod, sex, edad_al_ing_1, edad_ini_cons, escolaridad_rec, sus_principal_mod, freq_cons_sus_prin_ord, condicion_ocupacional_corr24, policonsumo, policonsumo2, num_hijos_mod_joel_bin, tenencia_de_la_vivienda_mod, macrozona, dg_cie_10_rec, dg_trs_cons_sus_or, clas_r, porc_pobr, sus_ini_mod_mvv, ano_nac_corr, con_quien_vive_joel, tipo_centro_pub, origen_ingreso_mod, compromiso_biopsicosocial,compromiso_biopsicosocial.y)))),] %>% nrow()/nrow(Base_fiscalia_v15f_grant_23_24_long2), accuracy = 0.1)
, "; total= ", format(nrow(Base_fiscalia_v15f_grant_23_24_long2),big.mark=","))[1] "Percentage of the total that has missing values: 84.4%; total= 90,075"Code
#[1] "Percentage of the total that has missing values: 84.4%; total= 90,075"
# 76029
# length(unique(Base_fiscalia_v15f_grant_23_24_long2$hash_key))
#patients: 72404
Base_fiscalia_v15f_grant_23_24_long %>% dplyr::filter(!motivodeegreso_mod_imp %in% c("En curso", "Muerte", "Derivación")) %>% distinct(hash_key) %>% nrow()[1] 72404Code
#[1] 72404
missing_pct <- colMeans(is.na(subset(Base_fiscalia_v15f_grant_23_24_long2, select= c(hash_key, tipo_de_plan_2, tipo_de_plan_2_mod, sex, edad_al_ing_1, edad_ini_cons, escolaridad_rec, sus_principal_mod, freq_cons_sus_prin_ord, condicion_ocupacional_corr24, policonsumo, policonsumo2, num_hijos_mod_joel_bin, tenencia_de_la_vivienda_mod, macrozona, dg_cie_10_rec, dg_trs_cons_sus_or, clas_r, porc_pobr, sus_ini_mod_mvv, ano_nac_corr, con_quien_vive_joel, tipo_centro_pub, origen_ingreso_mod, compromiso_biopsicosocial, compromiso_biopsicosocial.y
) ))) * 100
round(missing_pct,2) hash_key tipo_de_plan_2
0.00 0.10
tipo_de_plan_2_mod sex
0.12 0.00
edad_al_ing_1 edad_ini_cons
0.01 6.84
escolaridad_rec sus_principal_mod
0.43 0.00
freq_cons_sus_prin_ord condicion_ocupacional_corr24
0.50 0.00
policonsumo policonsumo2
0.00 0.00
num_hijos_mod_joel_bin tenencia_de_la_vivienda_mod
0.81 5.84
macrozona dg_cie_10_rec
0.02 0.00
dg_trs_cons_sus_or clas_r
0.00 0.00
porc_pobr sus_ini_mod_mvv
0.00 6.64
ano_nac_corr con_quien_vive_joel
0.01 0.00
tipo_centro_pub origen_ingreso_mod
0.02 0.00
compromiso_biopsicosocial compromiso_biopsicosocial.y
1.95 1.94 Code
set.seed(2125)
Base_fiscalia_v15f_grant_23_24_long2_miss <- Base_fiscalia_v15f_grant_23_24_long2 %>%
dplyr::arrange(hash_key, fech_ing_num) %>%
missRanger::missRanger(
formula = sus_principal_mod + condicion_ocupacional_corr24 + dg_trs_cons_sus_or + con_quien_vive_joel + clas_r + porc_pobr + edad_al_ing_1 + ano_nac_corr + macrozona + tipo_centro_pub + tipo_de_plan_2 + tipo_de_plan_2_mod + escolaridad_rec + freq_cons_sus_prin_ord + num_hijos_mod_joel_bin + compromiso_biopsicosocial.y + compromiso_biopsicosocial + tenencia_de_la_vivienda_mod + sus_ini_mod_mvv + edad_ini_cons ~ edad_al_ing_1 + sex + edad_ini_cons + escolaridad_rec + sus_principal_mod + freq_cons_sus_prin_ord + condicion_ocupacional_corr24 + policonsumo + num_hijos_mod_joel_bin + tenencia_de_la_vivienda_mod + macrozona + dg_cie_10_rec + dg_trs_cons_sus_or + clas_r + porc_pobr + sus_ini_mod_mvv + ano_nac_corr + con_quien_vive_joel + tipo_centro_pub + origen_ingreso_mod + compromiso_biopsicosocial + cens_time + policonsumo2 + compromiso_biopsicosocial.y + tipo_de_plan_2 + tipo_de_plan_2_mod,
num.trees = 300,
pmm.k = 5,
returnOOB=T,
maxiter= 50,
verbose = 2,
seed = 2125)
Missing value imputation by random forests
Variables to impute: sus_principal_mod, condicion_ocupacional_corr24, dg_trs_cons_sus_or, con_quien_vive_joel, clas_r, porc_pobr, edad_al_ing_1, ano_nac_corr, macrozona, tipo_centro_pub, tipo_de_plan_2, tipo_de_plan_2_mod, escolaridad_rec, freq_cons_sus_prin_ord, num_hijos_mod_joel_bin, compromiso_biopsicosocial.y, compromiso_biopsicosocial, tenencia_de_la_vivienda_mod, sus_ini_mod_mvv, edad_ini_cons
Variables used to impute: edad_al_ing_1, sex, edad_ini_cons, escolaridad_rec, sus_principal_mod, freq_cons_sus_prin_ord, condicion_ocupacional_corr24, policonsumo, num_hijos_mod_joel_bin, tenencia_de_la_vivienda_mod, macrozona, dg_cie_10_rec, dg_trs_cons_sus_or, clas_r, porc_pobr, sus_ini_mod_mvv, ano_nac_corr, con_quien_vive_joel, tipo_centro_pub, origen_ingreso_mod, compromiso_biopsicosocial, cens_time, policonsumo2, compromiso_biopsicosocial.y, tipo_de_plan_2, tipo_de_plan_2_mod
ss_pr_ cn__24 dg____ cn_q__ clas_r prc_pb ed___1 an_nc_ macrzn tp_cn_ tp___2 t___2_ esclr_ fr____ nm____ cmpr_. cmprm_ tn____ ss_n__ edd_n_
iter 1: 0.4744 0.4529 0.2608 0.4665 0.1676 0.5058 0.6236 0.0106 0.1195 0.1534 0.3502 0.0965 0.3121 0.3646 0.6032 0.3099 0.0592 0.3085 0.2574 0.5780
iter 2: 1e-04 0.0000 0.0000 0.0000 0.0000 0.0084 0.0013 0.0011 0.0000 0.0000 0.0000 0.0000 0.0000 1e-04 0.0012 0.0000 0.0000 4e-04 1e-04 0.0186
iter 3: 1e-04 0.0000 0.0000 0.0000 0.0000 0.0083 0.0012 0.0012 0.0000 0.0000 0.0000 0.0000 0.0000 1e-04 0.0011 0.0000 0.0000 5e-04 1e-04 0.0188 Code
invisible("2024-04-25: ! Can't subset columns that don't exist; Column `vars_cov` doesn't exist. I corrected that by including explicitly the covariates")
library(arrow)
library(rio)
arrow::write_parquet(Base_fiscalia_v15f_grant_23_24_long2_miss,
"_proposal_grant/2023/Base_fiscalia_v15f_grant_23_24_long2_miss240425.gz.parquet",
compression = "gzip", compression_level = 5)
column_names <- c(
"sus_principal_mod", "condicion_ocupacional_corr24", "dg_trs_cons_sus_or", "con_quien_vive_joel",
"clas_r", "porc_pobr", "edad_al_ing_1", "ano_nac_corr", "macrozona", "tipo_centro_pub",
"tipo_de_plan_2", "tipo_de_plan_2_mod", "escolaridad_rec", "freq_cons_sus_prin_ord",
"num_hijos_mod_joel_bin", "compromiso_biopsicosocial.y", "compromiso_biopsicosocial",
"tenencia_de_la_vivienda_mod", "sus_ini_mod_mvv", "edad_ini_cons"
)
rmses_reduction <- matrix(
c(
0.4744, 0.4529, 0.2608, 0.4665, 0.1676, 0.5058, 0.6236, 0.0106, 0.1195, 0.1534, 0.3502, 0.0965, 0.3121, 0.3646, 0.6032, 0.3099, 0.0592, 0.3085, 0.2574, 0.5780, 1e-04, 0.0000, 0.0000, 0.0000, 0.0000, 0.0084, 0.0013, 0.0011, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 1e-04, 0.0012, 0.0000, 0.0000, 4e-04, 1e-04, 0.0186, 1e-04, 0.0000, 0.0000, 0.0000, 0.0000, 0.0083, 0.0012, 0.0012, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 1e-04, 0.0011, 0.0000, 0.0000, 5e-04, 1e-04, 0.0188
),
nrow = 3, byrow = TRUE
)
rmses_reduction <- data.frame(rmses_reduction)
colnames(rmses_reduction) <- column_names
#dput(round(attr(Base_fiscalia_v15f_grant_23_24_long2_miss,"oob"),3))
print("Out of bag error:")[1] "Out of bag error:"Code
c(sus_principal_mod = 0,
condicion_ocupacional_corr24 = 0,
dg_trs_cons_sus_or = 0,
con_quien_vive_joel = 0,
clas_r = 0,
porc_pobr = 0.008,
edad_al_ing_1 = 0.001,
ano_nac_corr = 0.001,
macrozona = 0,
tipo_centro_pub = 0,
tipo_de_plan_2 = 0, tipo_de_plan_2_mod = 0,
escolaridad_rec = 0,
freq_cons_sus_prin_ord = 0,
num_hijos_mod_joel_bin = 0.001,
compromiso_biopsicosocial.y = 0,
compromiso_biopsicosocial = 0,
tenencia_de_la_vivienda_mod = 0,
sus_ini_mod_mvv = 0,
edad_ini_cons = 0.019) sus_principal_mod condicion_ocupacional_corr24
0.000 0.000
dg_trs_cons_sus_or con_quien_vive_joel
0.000 0.000
clas_r porc_pobr
0.000 0.008
edad_al_ing_1 ano_nac_corr
0.001 0.001
macrozona tipo_centro_pub
0.000 0.000
tipo_de_plan_2 tipo_de_plan_2_mod
0.000 0.000
escolaridad_rec freq_cons_sus_prin_ord
0.000 0.000
num_hijos_mod_joel_bin compromiso_biopsicosocial.y
0.001 0.000
compromiso_biopsicosocial tenencia_de_la_vivienda_mod
0.000 0.000
sus_ini_mod_mvv edad_ini_cons
0.000 0.019 Code
#OOB prediction error per iteration and variable (1 minus R-squared for regression)
#The default mtry in missRanger is sqrt(p), where p is the number of variables in the dataset.
#OOB prediction errors are quantified as 1 - R^2 for numeric variables, and as classification error otherwise. If a variable has been imputed only univariately, the value is 1.
#https://rdrr.io/cran/missRanger/man/missRanger.htmlIrreLong
Irregularity
We first formatted the data according what is being needed by the method used here: id which represents hash_keys in a numeric factored format, and generated the variable less90d_tr which is a variable that groups days in treatment between 0-90 in 1, and 0 for anything else. Additionally, we renamed the variable of biopsychosocial compromise longitudinally measured ( compromiso_biopsicosocial.y) for comp_bpsc_y for ease of processing.
Code
Base_fiscalia_v15f_grant_23_24_long2_proc<-
Base_fiscalia_v15f_grant_23_24_long2%>% #dplyr::filter(event==1)%>%
dplyr::group_by(hash_key) %>%
dplyr::mutate(is_first_occurrence = if_else(row_number() == 1, 1, 0)) %>%
#para hacer ID
dplyr::ungroup() %>%
dplyr::mutate(id = cumsum(is_first_occurrence)) %>%
dplyr::select(-is_first_occurrence) %>%
dplyr::mutate(less_90d_tr= factor(if_else(dias_treat_imp_sin_na<=90 & dias_treat_imp_sin_na>0,1,0,NA_real_)))%>%
dplyr::rename("comp_bpsc_y"="compromiso_biopsicosocial.y") %>%
data.frame()
Base_fiscalia_v15f_grant_23_24_long2_miss_proc<-
Base_fiscalia_v15f_grant_23_24_long2_miss%>% #dplyr::filter(event==1)%>%
dplyr::group_by(hash_key) %>%
dplyr::mutate(is_first_occurrence = if_else(row_number() == 1, 1, 0)) %>%
#para hacer ID
dplyr::ungroup() %>%
dplyr::mutate(id = cumsum(is_first_occurrence)) %>%
dplyr::select(-is_first_occurrence) %>%
dplyr::mutate(less_90d_tr= factor(if_else(dias_treat_imp_sin_na<=90 & dias_treat_imp_sin_na>0,1,0,NA_real_)))%>%
dplyr::rename("comp_bpsc_y"="compromiso_biopsicosocial.y") %>%
data.frame()
invisible("parece que había que especificar bien las palabras, no permite data.tables")Next, we measured the extent of irregularity between observation times and gaps between observations. We replicated this analysis in the imputed and original databases, and using calendar time as a reference of time taken.
Code
invisible(" Measuring extent of irregularity")
counts <- extent.of.irregularity(Base_fiscalia_v15f_grant_23_24_long2_proc,
time="time",
id="id",
scheduledtimes=NULL, cutpoints=NULL,ncutpts=50,
maxfu=16*24, plot=F, legendx=30, legendy=0.8,
formula=Surv(time.lag,time,event)~1,tau=12*12)
counts_counts<-
matrix(c(0, 0.81609, 0.18391, 0.4476, 0.49165, 0.06075, 0.61607,
0.35612, 0.02781, 0.70566, 0.27893, 0.0154, 0.76138, 0.22912,
0.0095, 0.7994, 0.19428, 0.00632, 0.82702, 0.16847, 0.00451,
0.84794, 0.14878, 0.00327, 0.86434, 0.1332, 0.00245, 0.87765,
0.12038, 0.00197, 0.88848, 0.11, 0.00152, 0.8976, 0.10118, 0.00122,
0.90537, 0.09359, 0.00103, 0.91202, 0.08712, 0.00085, 0.91781,
0.08147, 0.00072, 0.92287, 0.07653, 6e-04, 0.92734, 0.07216,
5e-04, 0.93132, 0.06826, 0.00042, 0.9349, 0.06473, 0.00037, 0.93815,
0.06151, 0.00034, 0.94107, 0.05863, 3e-04, 0.94371, 0.05603,
0.00025, 0.94615, 0.05361, 0.00023, 0.94838, 0.05141, 0.00021,
0.95043, 0.04938, 0.00019, 0.95233, 0.0475, 0.00017, 0.95408,
0.04576, 0.00015, 0.95572, 0.04413, 0.00015, 0.95724, 0.04262,
0.00014, 0.95866, 0.04122, 0.00012, 0.95999, 0.03989, 0.00012,
0.96123, 0.03867, 1e-04, 0.9624, 0.0375, 1e-04, 0.9635, 0.03642,
8e-05, 0.96454, 0.03538, 8e-05, 0.96552, 0.03441, 7e-05, 0.96645,
0.03348, 7e-05, 0.96733, 0.03261, 7e-05, 0.96817, 0.03177, 6e-05,
0.96896, 0.03099, 6e-05, 0.96971, 0.03025, 5e-05, 0.97043, 0.02952,
5e-05, 0.97112, 0.02883, 5e-05, 0.97177, 0.02818, 4e-05, 0.9724,
0.02756, 4e-05, 0.973, 0.02696, 4e-05, 0.97357, 0.02639, 4e-05,
0.97412, 0.02584, 4e-05, 0.97464, 0.02533, 3e-05, 0.97515, 0.02482,
3e-05),
byrow=T,
nrow = 50, ncol= 3, dimnames = list(NULL, c("X1", "X2", "X3")))
counts$auc[1] 0.06548913Code
# $auc
# [1] 0.06548913
counts$transformed.auc[1] 20.25358Code
# $transformed.auc
# [1] 20.25358
invisible("Is not that much, but is more than 0")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
counts_miss <- extent.of.irregularity(Base_fiscalia_v15f_grant_23_24_long2_miss_proc,
time="time",
id="id",
scheduledtimes=NULL, cutpoints=NULL,ncutpts=50,
maxfu=16*24, plot=F, legendx=30, legendy=0.8,
formula=Surv(time.lag,time,event)~1,tau=12*12)
counts_miss_counts<-
matrix(c(0, 0.81609, 0.18391, 0.4476, 0.49165, 0.06075, 0.61607,
0.35612, 0.02781, 0.70566, 0.27893, 0.0154, 0.76138, 0.22912,
0.0095, 0.7994, 0.19428, 0.00632, 0.82702, 0.16847, 0.00451,
0.84794, 0.14878, 0.00327, 0.86434, 0.1332, 0.00245, 0.87765,
0.12038, 0.00197, 0.88848, 0.11, 0.00152, 0.8976, 0.10118, 0.00122,
0.90537, 0.09359, 0.00103, 0.91202, 0.08712, 0.00085, 0.91781,
0.08147, 0.00072, 0.92287, 0.07653, 6e-04, 0.92734, 0.07216,
5e-04, 0.93132, 0.06826, 0.00042, 0.9349, 0.06473, 0.00037, 0.93815,
0.06151, 0.00034, 0.94107, 0.05863, 3e-04, 0.94371, 0.05603,
0.00025, 0.94615, 0.05361, 0.00023, 0.94838, 0.05141, 0.00021,
0.95043, 0.04938, 0.00019, 0.95233, 0.0475, 0.00017, 0.95408,
0.04576, 0.00015, 0.95572, 0.04413, 0.00015, 0.95724, 0.04262,
0.00014, 0.95866, 0.04122, 0.00012, 0.95999, 0.03989, 0.00012,
0.96123, 0.03867, 1e-04, 0.9624, 0.0375, 1e-04, 0.9635, 0.03642,
8e-05, 0.96454, 0.03538, 8e-05, 0.96552, 0.03441, 7e-05, 0.96645,
0.03348, 7e-05, 0.96733, 0.03261, 7e-05, 0.96817, 0.03177, 6e-05,
0.96896, 0.03099, 6e-05, 0.96971, 0.03025, 5e-05, 0.97043, 0.02952,
5e-05, 0.97112, 0.02883, 5e-05, 0.97177, 0.02818, 4e-05, 0.9724,
0.02756, 4e-05, 0.973, 0.02696, 4e-05, 0.97357, 0.02639, 4e-05,
0.97412, 0.02584, 4e-05, 0.97464, 0.02533, 3e-05, 0.97515, 0.02482,
3e-05),
# 2024-03-07: i realized that i retrieved the information wrong so i need the byrow
byrow=T,
nrow = 50, ncol= 3, dimnames = list(NULL, c("X1", "X2", "X3")))
counts_miss$auc[1] 0.06548913Code
#[1] 0.06548913
counts_miss$transformed.auc[1] 20.25358Code
#[1] 20.25358
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("2024-04-27: restrict to people w/ >1 treatment")
Base_fiscalia_v15f_grant_23_24_long2_miss_proc_multtr<-
Base_fiscalia_v15f_grant_23_24_long2_miss_proc %>%
dplyr::group_by(hash_key) %>% dplyr::mutate(n=n()) %>% dplyr::mutate(is_first_occurrence = if_else(row_number() == 1, 1, 0)) %>% dplyr::ungroup() %>% dplyr::filter(case_when(n>1~ T,T~F)) %>% dplyr::mutate(id = factor(cumsum(is_first_occurrence))) %>% data.frame()
#dplyr::filter(max_treatment>1)
paste0("p= ", length(unique(Base_fiscalia_v15f_grant_23_24_long2_miss_proc_multtr$hash_key)),
"; n= ", nrow(Base_fiscalia_v15f_grant_23_24_long2_miss_proc_multtr))[1] "p= 13317; n= 30988"Code
counts_restricted <- extent.of.irregularity(Base_fiscalia_v15f_grant_23_24_long2_miss_proc_multtr,
time="time",
id="id",
scheduledtimes=NULL, cutpoints=NULL,ncutpts=50,
maxfu=16*24, plot=F, legendx=30, legendy=0.8,
formula=Surv(time.lag,time,event)~1,tau=12*12)
#counts_restricted$counts %>% round(.,5) %>% dput()
counts_restricted_counts<-
matrix(c(0, 8e-05, 0.99992, 0.21724, 0.45048, 0.33228, 0.39361,
0.45328, 0.15311, 0.51048, 0.40435, 0.08517, 0.59123, 0.35607,
0.0527, 0.64965, 0.31514, 0.03522, 0.69401, 0.28082, 0.02517,
0.7286, 0.25287, 0.01853, 0.75589, 0.23032, 0.01379, 0.77875,
0.21026, 0.01099, 0.79751, 0.19379, 0.0087, 0.81332, 0.17973,
0.00695, 0.8271, 0.16702, 0.00588, 0.83873, 0.1565, 0.00477,
0.84907, 0.14687, 0.00405, 0.85816, 0.13838, 0.00347, 0.86615,
0.13091, 0.00294, 0.87323, 0.12434, 0.00243, 0.87973, 0.11814,
0.00213, 0.88562, 0.11247, 0.00191, 0.89094, 0.10736, 0.0017,
0.89575, 0.10278, 0.00147, 0.90021, 0.09844, 0.00135, 0.90428,
0.09453, 0.0012, 0.90803, 0.09088, 0.00109, 0.91152, 0.08749,
0.00099, 0.91473, 0.08438, 0.00089, 0.91773, 0.08145, 0.00082,
0.92056, 0.07867, 0.00077, 0.92317, 0.07612, 0.00071, 0.92562,
0.07371, 0.00067, 0.92791, 0.07148, 0.00061, 0.93006, 0.06938,
0.00056, 0.93207, 0.06743, 5e-04, 0.93399, 0.06556, 0.00046,
0.93578, 0.06382, 4e-04, 0.93752, 0.06209, 4e-04, 0.93916, 0.06046,
0.00038, 0.94071, 0.05892, 0.00037, 0.94216, 0.05751, 0.00033,
0.94356, 0.05613, 0.00031, 0.94491, 0.05478, 0.00031, 0.94617,
0.05356, 0.00028, 0.94736, 0.05239, 0.00025, 0.94853, 0.05123,
0.00024, 0.94965, 0.05012, 0.00023, 0.95072, 0.04906, 0.00022,
0.95173, 0.04806, 2e-04, 0.95271, 0.04709, 0.00019, 0.95365,
0.04617, 0.00018),
# 2024-03-07: i realized that i retrieved the information wrong so i need the byrow
byrow=T,
nrow = 50, ncol= 3, dimnames = list(NULL, c("X1", "X2", "X3")))
counts_restricted$auc[1] 0.212881Code
#[1] 0.213
counts_restricted$transformed.auc[1] 80.02791Code
#80
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Check if there were any errors in of lag times after time")
error_in_time_followup<-
Base_fiscalia_v15f_grant_23_24_long2_proc %>%
dplyr::group_by(hash_key) %>%
dplyr::mutate(lag_time= lag(time)) %>%
dplyr::ungroup() %>%
dplyr::filter(dplyr::case_when(time <=lag_time & !is.na(lag_time)~T,T~F)) %>%
dplyr::select(hash_key, time, lag_time, motivodeegreso_mod_imp, dias_treat_imp_sin_na, enter_time, fech_ing_num, fech_egres_num, cens_time)
# A tibble: 0 x 9
Base_fiscalia_v15f_grant_23_24_long2_proc %>%
dplyr::group_by(hash_key) %>%
dplyr::mutate(lag_time= lag(time)) %>%
dplyr::ungroup() %>%
dplyr::filter(hash_key %in% unique(error_in_time_followup$hash_key)) %>%
dplyr::select(hash_key, time, motivodeegreso_mod_imp, dias_treat_imp_sin_na, enter_time, fech_ing_num, fech_egres_num, cens_time)# A tibble: 0 x 8
# ... with 8 variables: hash_key <chr>, time <dbl>,
# motivodeegreso_mod_imp <chr>, dias_treat_imp_sin_na <dbl>,
# enter_time <dbl>, fech_ing_num <dbl>, fech_egres_num <dbl>, cens_time <dbl>Code
# A tibble: 0 x 8
error_in_time_followup_miss<-
Base_fiscalia_v15f_grant_23_24_long2_miss_proc %>%
dplyr::group_by(hash_key) %>%
dplyr::mutate(lag_time= lag(time)) %>%
dplyr::ungroup() %>%
dplyr::filter(dplyr::case_when(time <=lag_time & !is.na(lag_time)~T,T~F)) %>%
dplyr::select(hash_key, time, lag_time, motivodeegreso_mod_imp, dias_treat_imp_sin_na, enter_time, fech_ing_num, fech_egres_num, cens_time)
# A tibble: 0 x 9
Base_fiscalia_v15f_grant_23_24_long2_miss_proc %>%
dplyr::group_by(hash_key) %>%
dplyr::mutate(lag_time= lag(time)) %>%
dplyr::ungroup() %>%
dplyr::filter(hash_key %in% unique(error_in_time_followup_miss$hash_key)) %>%
dplyr::select(hash_key, time, motivodeegreso_mod_imp, dias_treat_imp_sin_na, enter_time, fech_ing_num, fech_egres_num, cens_time) %>%
{
if (nrow(.) > 0) {
stop("There are ties or previous events that occured later than the actual events")
} else {
# Behavior when the condition isn't met (e.g., return an empty data frame)
data.frame(.)
}
}[1] hash_key time motivodeegreso_mod_imp
[4] dias_treat_imp_sin_na enter_time fech_ing_num
[7] fech_egres_num cens_time
<0 rows> (or 0-length row.names)Code
# [1] hash_key time motivodeegreso_mod_imp dias_treat_imp_sin_na enter_time
# [6] fech_ing_num fech_egres_num cens_time
# <0 rows> (or 0-length row.names)
invisible( "There are no ties or previous events that occured later than the following events")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Now we do it with calendar date unaltered")
set.seed(2125)
counts_corr <- extent.of.irregularity(dplyr::mutate(Base_fiscalia_v15f_grant_23_24_long2_proc, fech_egres_num= fech_egres_num/30.5),
time="fech_egres_num",
id="id",
scheduledtimes=NULL, cutpoints=NULL,ncutpts=50,
maxfu=18213/30.5, plot=F, legendx=30, legendy=0.8,
formula=Surv(fech_egres_num.lag,fech_egres_num,event)~1,tau=12*12)
counts_corr_counts<-
matrix(c(0.00023, 0.81592, 0.18384, 0.46307, 0.46597, 0.07096,
0.62881, 0.33313, 0.03807, 0.71395, 0.26336, 0.02269, 0.76724,
0.21789, 0.01487, 0.80367, 0.18596, 0.01037, 0.8302, 0.16231,
0.00749, 0.85036, 0.14405, 0.00559, 0.86621, 0.12954, 0.00425,
0.87903, 0.11765, 0.00332, 0.88967, 0.10767, 0.00266, 0.89855,
0.09931, 0.00214, 0.90617, 0.09204, 0.00179, 0.91267, 0.08586,
0.00147, 0.91832, 0.08047, 0.00121, 0.92335, 0.07558, 0.00107,
0.92776, 0.07134, 9e-04, 0.93165, 0.06762, 0.00073, 0.93519,
0.06417, 0.00064, 0.93839, 0.06104, 0.00057, 0.94128, 0.05822,
5e-04, 0.94391, 0.05566, 0.00043, 0.94631, 0.05332, 0.00038,
0.94852, 0.05115, 0.00034, 0.95054, 0.04917, 0.00029, 0.95245,
0.04727, 0.00028, 0.95418, 0.04558, 0.00024, 0.95581, 0.04398,
0.00022, 0.9573, 0.04251, 0.00019, 0.95873, 0.04108, 0.00019,
0.96004, 0.0398, 0.00016, 0.96129, 0.03855, 0.00016, 0.96245,
0.03741, 0.00014, 0.96355, 0.03633, 0.00013, 0.96458, 0.0353,
0.00012, 0.96555, 0.03435, 1e-04, 0.96649, 0.03342, 1e-04, 0.96737,
0.03254, 9e-05, 0.96819, 0.03172, 8e-05, 0.96898, 0.03094, 8e-05,
0.96974, 0.03019, 7e-05, 0.97046, 0.02948, 7e-05, 0.97113, 0.02881,
5e-05, 0.97179, 0.02815, 6e-05, 0.97242, 0.02753, 5e-05, 0.97301,
0.02694, 5e-05, 0.97358, 0.02638, 4e-05, 0.97413, 0.02583, 4e-05,
0.97466, 0.0253, 4e-05, 0.97516, 0.0248, 4e-05),
nrow = 50, ncol= 3,
# i realized that i needed the byrow=T because i used bad the reading of the dput
byrow=T, dimnames = list(NULL, c("X1", "X2", "X3")))
counts_corr$auc[1] 0.07266227Code
# $auc
# [1] 0.07266227
#
counts_corr$transformed.auc[1] 22.65514Code
# $transformed.auc
# [1] 22.65514
if(exists("omitted")){
invisible("To check specific gap bin widths for the 3 options in a specific time after 'filter'==1 (e.g.)")
rbind.data.frame(cbind.data.frame(type=rep("1. Modified",), time= 1:50, counts_counts),cbind.data.frame(type=rep("2. Modified\n(imputation)",), time= 1:50, counts_miss_counts), cbind.data.frame(type=rep("3. More 1\ntreatment",), time= 1:50, counts_restricted_counts), cbind.data.frame(type=rep("4. Original\ncalendar",), time= 1:50, counts_corr_counts)) %>% dplyr::filter(time==2) %>% dplyr::mutate_if(is.numeric,~round(.,2))
rbind.data.frame(cbind.data.frame(type=rep("1. Modified",), time= 1:50, counts_counts),cbind.data.frame(type=rep("2. Modified\n(imputation)",), time= 1:50, counts_miss_counts), cbind.data.frame(type=rep("3. More 1\ntreatment",), time= 1:50, counts_restricted_counts), cbind.data.frame(type=rep("4. Original\ncalendar",), time= 1:50, counts_corr_counts)) %>% dplyr::filter(time==6) %>% dplyr::mutate_if(is.numeric,~round(.,2))
#2024-05-13
cbind.data.frame(type=rep("5. Restricted\ntreatments",), time= 1:50, counts_restricted_counts) %>% dplyr::filter(time==2) %>% dplyr::mutate_if(is.numeric,~round(.,2))
cbind.data.frame(type=rep("5. Restricted\ntreatments",), time= 1:50, counts_restricted_counts) %>% dplyr::filter(time==6) %>% dplyr::mutate_if(is.numeric,~round(.,2))
}
plot_irregularity<-
rbind.data.frame(cbind.data.frame(type=rep("1. Modified",), time= 1:50, counts_counts),cbind.data.frame(type=rep("2. Modified\n(imputation)",), time= 1:50, counts_miss_counts),cbind.data.frame(type=rep("3. More 1\ntreatment",), time= 1:50, counts_restricted_counts),cbind.data.frame(type=rep("4. Original\ncalendar",), time= 1:50, counts_corr_counts)) %>%
reshape2::melt(id=c("type", "time")) %>%
dplyr::mutate(Observations= factor(variable, labels=c("0 obs per bin","1 obs per bin", "2+ obs per bin"))) %>%
ggplot(aes(x=time, y=value, linetype=Observations), linewidth=1)+
geom_line()+
facet_wrap(~type)+ #color=factor(type))
theme_classic()+
xlab("Bin width (%)")+
ylab("Mean proportions of individuals")#+
# scale_linetype_manual(name="Observations")
plot_irregularity
Code
plot_irregularity_05_2024<-
rbind.data.frame(cbind.data.frame(type=rep("3. More 1\ntreatment",), time= 1:50, counts_restricted_counts)) %>%
reshape2::melt(id=c("type", "time")) %>%
dplyr::mutate(Observations= factor(variable, labels=c("0 obs per bin","1 obs per bin", "2+ obs per bin"))) %>%
ggplot(aes(x=time, y=value, linetype=Observations), linewidth=1)+
geom_line()+
#facet_wrap(~type)+ #color=factor(type))
theme_classic()+
xlab("Bin width (%)")+
ylab("Mean proportions of individuals")#+
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
ggsave(plot_irregularity, file="_proposal_grant/2023/_figs/irreg240427.jpg", width=8.5, height=5.5, dpi=500)
ggsave(plot_irregularity_05_2024, file= paste0(folder_path,"_figs/irreg240513.jpg"), width=8.5, height=5.5, dpi=500)The AUCs that suggest that although low values, we can suspect a share of a counting process behind visits rather than a perfect repeated measures. Thus, it seems reasonable to conduct an analysis that consider irregularity in observations and an extent of informative assessment times, conditional on a bulk of covariates, including past observed outcomes, past assessment history, and baseline covariates. The graphic show the mean proportions of individuals with 0,1, and >1 visits per bin as bin width varies from 1 to 50% of the gap between admissions. In the graphic, we can see that in the first 5 bins there were more than one observation in the same bin of time.
IIW
Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#load("C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/an_grant_23_24_04_15.RData")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
# formulaph: the formula for the proportional hazards model for the visit intensity that will be used to derive inverse-intensity weights. The formula should usually use the counting process format (i.e. Surv(start,stop,event)). If a frailty model is used, the cluster(id) term should appear before other covariates
#
# formulanull: if stabilised weights are to be used, the formula for the null model used to stabilise the weights
#
# data: data frame containing the variables in the model
#
# id: character string indicating which column of the data identifies subjects
#
# time: character string indicating which column of the data contains the time at which the visit occurred
#
# event: character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
#
# lagvars: a vector of variable names corresponding to variables which need to be lagged by one visit to fit the visit intensity model. Typically time will be one of these variables. The function will internally add columns to the data containing the values of the lagged variables from the previous visit. Values of lagged variables for a subject's first visit will be set to NA. To access these variables in specifying the proportional hazards formulae, add ".lag" to the variable you wish to lag. For example, if time is the variable for time, time.lag is the time of the previous visit
#
# invariant: a vector of variable names corresponding to variables in data that are time-invariant. It is not necessary to list every such variable, just those that are invariant and also included in the proportional hazards model
#
# maxfu: the maximum follow-up time(s). If everyone is followed for the same length of time, this can be given as a single value. If individuals have different follow-up times, maxfu should have the same number of elements as there are rows of data
#
# lagfirst: A vector giving the value of each lagged variable for the first time within each subject. This is helpful if, for example, time is the variable to be lagged and you know that all subjects entered the study at time zero
#
# first: logical variable. If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred; in this case the baseline visit is observed with probability 1.
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("2024-03-06: No es necesario correr el proceso longitudinal con ...")
invisible("fech_ing_num.lag,fech_ing_num, porque ya corregi time y no debiese tener ...")
invisible("superposiciones con lag.time")
invisible("Por lo visto si hay un problema, pq no recupero ninguno")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("2014-03-08: Primero tenemos que ver si es plausible ver evidencia en contra de eventos aleatorios. ")
# glimpse(Phenobarb)
invisible("Correct id so it resemble Phenobarb")
if(class(Base_fiscalia_v15f_grant_23_24_long2_miss_proc$id)!="factor"){
Base_fiscalia_v15f_grant_23_24_long2_miss_proc$id <- factor(Base_fiscalia_v15f_grant_23_24_long2_miss_proc$id)
}
invisible("Correct id so it resemble Phenobarb")
data_mine_miss <- Base_fiscalia_v15f_grant_23_24_long2_miss_proc
invisible("Correct id so it resemble Phenobarb")
if(class(Base_fiscalia_v15f_grant_23_24_long2_miss_proc_multtr$id)!="factor"){
Base_fiscalia_v15f_grant_23_24_long2_miss_proc_multtr$id <- factor(Base_fiscalia_v15f_grant_23_24_long2_miss_proc_multtr$id)
}
data_mine_miss_restr <- Base_fiscalia_v15f_grant_23_24_long2_miss_proc_multtr
invisible("2024-04-15: Need to take into account that this implies having only those who experiment the event, which might be bad. I have to have people that is observed. This should be the event")
# BAD: data_mine_miss <- data_mine_miss[data_mine_miss$event==1,]
invisible("2024-04-15: keep the tr. outcome")
data_mine_miss$tr_outcome <- data_mine_miss$event
data_mine_miss_restr$tr_outcome <- data_mine_miss_restr$event
invisible("2024-04-15: transform the event into 1 (being observed")
data_mine_miss$event<-1
data_mine_miss_restr$event<-1
data_mine_miss_proc <-
cbind.data.frame(data_mine_miss,model.matrix(~ comp_bpsc_y - 1 + less_90d_tr -1, data = data.frame(data_mine_miss))) %>% dplyr::rename("comp_bpsc_y2_moderate"="comp_bpsc_y2-Moderate","comp_bpsc_y3_severe"="comp_bpsc_y3-Severe") %>% data.frame()
data_mine_miss_restr_proc <-
cbind.data.frame(data_mine_miss_restr,model.matrix(~ comp_bpsc_y - 1 + less_90d_tr -1, data = data.frame(data_mine_miss_restr))) %>% dplyr::rename("comp_bpsc_y2_moderate"="comp_bpsc_y2-Moderate","comp_bpsc_y3_severe"="comp_bpsc_y3-Severe") %>% data.frame()
invisible("2024-04-18: eliminate Mild level")
data_mine_miss_proc$comp_bpsc_y1.Mild<-NULL
data_mine_miss_restr_proc$comp_bpsc_y1.Mild<-NULL
invisible("2024-04-15: following the addcensoredrows function vignette [https://search.r-project.org/CRAN/refmans/IrregLong/html/addcensoredrows.html] we crated a database to compute")
maxfu_df<-
data_mine_miss_proc %>%
dplyr::group_by(id) %>%
dplyr::summarise(maxfu.time= max(cens_time, na.rm=T)) %>% #,0)
dplyr::rename("maxfu.id"="id") %>%
as.data.frame()
maxfu_restr_df<-
data_mine_miss_restr_proc %>%
dplyr::group_by(id) %>%
dplyr::summarise(maxfu.time= max(cens_time, na.rm=T)) %>% #,0)
dplyr::rename("maxfu.id"="id") %>%
as.data.frame()
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("2024-04-15: Make a sample of 10% to test quicker")
data_mine_miss_proc_sample<-
data_mine_miss_proc %>%
dplyr::filter(id %in%
as.character(sample(1:length(unique(data_mine_miss_proc$id)),5e3)))
maxfu_sample_df<-
data_mine_miss_proc_sample %>%
dplyr::group_by(id) %>%
dplyr::summarise(maxfu.time= round(max(cens_time, na.rm=T),0)) %>%
dplyr::rename("maxfu.id"="id") %>%
as.data.frame()
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("i0: only events; missing values, only event.lag")
i0 <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag +
cluster(id),
id= "id",
time= "time",
event= "event", #event: dropout; 2024-04-18 being reobserved (visit)
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome"),#"less_90d_tr", "comp_bpsc_y"), #
maxfu= maxfu_df,#12*12,# Base_fiscalia_v15f_grant_23_24_long2$cens_time, #12*12,
lagfirst= c(0,0),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
i0$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag,
data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z p
tr_outcome.lag -1.868626 0.154336 0.009581 0.009747 -191.7 <2e-16
Likelihood ratio test=50186 on 1 df, p=< 2.2e-16
n= 162323, number of events= 89925
(150 observations deleted due to missingness)Code
broom::tidy(i0$m)# A tibble: 1 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag -1.87 0.00958 0.00975 -192. 0Code
#broom::tidy(i0$m) %>% dput()
i0_df_out<-
cbind.data.frame(term = "tr_outcome.lag", estimate = -1.86862564229299,
std.error = 0.00958080874903184, robust.se = 0.00974694287151046,
statistic = -191.714024276765, p.value = 0)%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
i0$datacox %>%
dplyr::group_by(id) %>%
dplyr::mutate(lag_time= lag(time)) %>%
dplyr::ungroup() %>%
dplyr::filter(dplyr::case_when(time <=lag_time & !is.na(lag_time)~T,T~F)) %>%
dplyr::select(hash_key, time, lag_time, motivodeegreso_mod_imp, dias_treat_imp_sin_na, enter_time, fech_ing_num, fech_egres_num, cens_time)%>%
{
if (nrow(.) > 0) {
stop("There are ties or previous events that occured later than the actual events")
} else {
# Behavior when the condition isn't met (e.g., return an empty data frame)
data.frame(.)
}
}[1] hash_key time lag_time
[4] motivodeegreso_mod_imp dias_treat_imp_sin_na enter_time
[7] fech_ing_num fech_egres_num cens_time
<0 rows> (or 0-length row.names)Code
invisible("it seems that there are no errors")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("i0b: only events; missing values, only event.lag; lag first 1s")
i0b <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag + #comp_bpsc_y.lag+ less_90d_tr.lag+
cluster(id),
id= "id",
time= "time",
event= "event", #event: dropout
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome"),#"less_90d_tr", "comp_bpsc_y"), #
maxfu= maxfu_df,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(1,1),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
i0b$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag,
data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z p
tr_outcome.lag 1.06837 2.91063 0.01705 0.01690 63.21 <2e-16
Likelihood ratio test=5298 on 1 df, p=< 2.2e-16
n= 159556, number of events= 87158
(2917 observations deleted due to missingness)Code
broom::tidy(i0b$m)# A tibble: 1 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 1.07 0.0170 0.0169 63.2 0Code
i0b_df_out<-
cbind.data.frame(term = "tr_outcome.lag", estimate = 1.06836854522169,
std.error = 0.0170488969088114, robust.se = 0.0169029487257102,
statistic = 63.2060454396721, p.value = 0)%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i0$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
1.000 1.000 1.000 1.846 1.000 6.479 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 1.000 1.000 1.000 1.846 1.000 6.479
ploti0<-
ggplot( data= as.data.frame(i0$iiw.weight),aes(x= `i0$iiw.weight`))+
geom_histogram(bins=30)+
theme_classic()+ xlim(0,10)
#Removed 13708 rows containing non-finite values (`stat_bin()`).
summary(i0b$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.3436 1.0000 1.0000 0.8986 1.0000 1.0000 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.3436 1.0000 1.0000 0.8986 1.0000 1.0000
ploti0b<-
ggplot( data= as.data.frame(i0b$iiw.weight),aes(x= `i0b$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Restrict to users with more than one treatment")
data_mine_miss_proc_multtr<-
data_mine_miss_proc %>%
dplyr::group_by(hash_key) %>% dplyr::mutate(n=n()) %>% dplyr::ungroup() %>% dplyr::filter(case_when(n>1~ T,T~F)) %>% data.frame()
#dplyr::filter(max_treatment>1)
invisible("2014-04-15: make a follow-up for this subsample")
maxfu_multtr_df<-
data_mine_miss_proc_multtr %>%
dplyr::group_by(id) %>%
dplyr::summarise(maxfu.time= max(cens_time, na.rm=T)) %>% #,0)
dplyr::rename("maxfu.id"="id") %>%
as.data.frame()
invisible("Remember that his has only events (observations)")
paste0("The original database had ",
format(length(unique(data_mine_miss_proc$hash_key)), big.mark=","),
" patients and ",
format(length(data_mine_miss_proc$hash_key), big.mark=","),
" observations with only events, but keeping patients with more than one treatment left only ",
format(length(unique(data_mine_miss_proc_multtr$hash_key)), big.mark=","),
" patients and ",
format(length(data_mine_miss_proc_multtr$hash_key), big.mark=","),
" observations"
)[1] "The original database had 72,404 patients and 90,075 observations with only events, but keeping patients with more than one treatment left only 13,317 patients and 30,988 observations"Code
#[1] "The original database had 72,404 patients and 90,075 observations with only events, but keeping patients with more than one treatment left only 13,317 patients and 30,988 observations"
invisible("multiple treatments; only event.lag; lag first 0s")
i0c <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag + #comp_bpsc_y.lag+ less_90d_tr.lag+
cluster(id),
id= "id",
time= "time",
event= "event", #event: dropout
data= data_mine_miss_proc_multtr,
invariant= "id",
lagvars= c("time", "tr_outcome"),#"less_90d_tr", "comp_bpsc_y"), #
maxfu= maxfu_multtr_df, #12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(0,0),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
# In Surv(fech_ing_num.lag, fech_ing_num, event) :
# Stop time must be > start time, NA created
i0c$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag,
data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z p
tr_outcome.lag -0.747183 0.473699 0.012291 0.008773 -85.17 <2e-16
Likelihood ratio test=3636 on 1 df, p=< 2.2e-16
n= 44272, number of events= 30959
(29 observations deleted due to missingness)Code
broom::tidy(i0c$m)# A tibble: 1 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag -0.747 0.0123 0.00877 -85.2 0Code
i0c_df_out<-
cbind.data.frame(term = "tr_outcome.lag", estimate = -0.747182521538917,
std.error = 0.0122906036735641, robust.se = 0.00877334474470237,
statistic = -85.1650702532908, p.value = 0)%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
i0d <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag + #comp_bpsc_y.lag+ less_90d_tr.lag+
cluster(id),
id= "id",
time= "time",
event= "event", #event: observed
data= data_mine_miss_proc_multtr,
invariant= "id",
lagvars= c("time", "tr_outcome"),#"less_90d_tr", "comp_bpsc_y"), #
maxfu= maxfu_multtr_df,#12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(1,1),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
# In Surv(fech_ing_num.lag, fech_ing_num, event) :
# Stop time must be > start time, NA created
i0d$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag,
data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z p
tr_outcome.lag 0.27465 1.31607 0.01798 0.01291 21.27 <2e-16
Likelihood ratio test=247.4 on 1 df, p=< 2.2e-16
n= 43638, number of events= 30325
(663 observations deleted due to missingness)Code
broom::tidy(i0d$m)# A tibble: 1 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 0.275 0.0180 0.0129 21.3 2.28e-100Code
i0d_df_out<-
cbind.data.frame(term = "tr_outcome.lag", estimate = 0.274653363170145,
std.error = 0.0179832241105765, robust.se = 0.0129143329186389,
statistic = 21.2673287037339, p.value = 2.27894355540563e-100)%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i0c$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
1.000 1.000 1.000 1.499 2.111 2.111 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 1.000 1.000 1.000 1.499 2.111 2.111
ploti0c<-
ggplot( data= as.data.frame(i0c$iiw.weight),aes(x= `i0c$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
summary(i0d$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.7598 0.7598 0.7598 0.7890 0.7598 1.0000 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.7598 0.7598 0.7598 0.7890 0.7598 1.0000
ploti0d<-
ggplot( data= as.data.frame(i0d$iiw.weight),aes(x= `i0d$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
invisible("Now we use the 'data_mine_miss_proc' and try to dummify previous events")
i1 <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ #less_90d_tr.lag+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1", "policonsumo2"),
maxfu= maxfu_df,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= rep(0),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
invisible("No longer Error in as.character.factor(x) : malformed factor")
# Además: Warning message:
# In Surv(time.lag, time, event) : Stop time must be > start time, NA created
invisible("Glimpse")
data_mine_miss_proc[,c("id", "time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1")] %>% glimpse()Rows: 90,075
Columns: 6
$ id <fct> 1, 2, 3, 4, 5, 6, 7, 7, 8, 9, 10, 11, 12, 13, 14~
$ time <dbl> 9.8032787, 8.5901639, 8.8196721, 1.1475410, 7.01~
$ tr_outcome <dbl> 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, ~
$ comp_bpsc_y2_moderate <dbl> 0, 1, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 1, 1, ~
$ comp_bpsc_y3_severe <dbl> 1, 0, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 0, ~
$ less_90d_tr1 <dbl> 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, ~Code
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
i1b <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ #less_90d_tr.lag+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2"),
maxfu= maxfu_df,#12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(1,1,0,1,1,1),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
i1$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag, data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z p
tr_outcome.lag 0.17918 1.19624 0.01920 0.01968 9.105 < 2e-16
comp_bpsc_y2_moderate.lag 0.15399 1.16648 0.03195 0.03228 4.771 1.83e-06
comp_bpsc_y3_severe.lag 0.52003 1.68207 0.03238 0.03320 15.662 < 2e-16
less_90d_tr1.lag 0.04020 1.04102 0.01800 0.01847 2.176 0.0295
policonsumo2.lag 0.20122 1.22289 0.01985 0.02087 9.640 < 2e-16
Likelihood ratio test=990.1 on 5 df, p=< 2.2e-16
n= 90069, number of events= 17671
(72404 observations deleted due to missingness)Code
broom::tidy(i1$m)# A tibble: 5 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 0.179 0.0192 0.0197 9.10 8.65e-20
2 comp_bpsc_y2_moderate.lag 0.154 0.0320 0.0323 4.77 1.83e- 6
3 comp_bpsc_y3_severe.lag 0.520 0.0324 0.0332 15.7 2.74e-55
4 less_90d_tr1.lag 0.0402 0.0180 0.0185 2.18 2.95e- 2
5 policonsumo2.lag 0.201 0.0199 0.0209 9.64 5.40e-22Code
i1_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag",
"comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag"
), estimate = c(0.179303997546012, 0.155159976117794, 0.522643191348996,
0.0396946478038892, 0.200990917537315), std.error = c(0.0192034667885392,
0.0319537357893297, 0.0323780221201328, 0.0180015312824529, 0.0198507378783121
), robust.se = c(0.0196804675715693, 0.0322870898718578, 0.0332167392234256,
0.0184789094834668, 0.020873477226514), statistic = c(9.11075902510759,
4.80563521623034, 15.7343316522896, 2.14810553833841, 9.62900983656
), p.value = c(8.18076035588734e-20, 1.54260938373677e-06, 8.79807768652411e-56,
0.0317053723627336, 6.03077841071276e-22))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
invisible("It is not necessary to investigate violations of the counting process (progressive times)")
invisible("Does really PSU in the previous treatment protects?. 2024-04-15: now it make sense, every coeff is HR>1!")
i1b$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag, data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z p
tr_outcome.lag 0.204432 1.226827 0.018508 0.020385 10.029 <2e-16
comp_bpsc_y2_moderate.lag 0.002844 1.002849 0.031859 0.033560 0.085 0.932
comp_bpsc_y3_severe.lag 1.435238 4.200644 0.030608 0.031895 44.999 <2e-16
less_90d_tr1.lag 1.279551 3.595026 0.011541 0.011892 107.597 <2e-16
policonsumo2.lag 0.840131 2.316670 0.018508 0.021128 39.765 <2e-16
Likelihood ratio test=61818 on 5 df, p=< 2.2e-16
n= 159556, number of events= 87158
(2917 observations deleted due to missingness)Code
broom::tidy(i1b$m)# A tibble: 5 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 0.204 0.0185 0.0204 10.0 1.14e-23
2 comp_bpsc_y2_moderate.lag 0.00284 0.0319 0.0336 0.0848 9.32e- 1
3 comp_bpsc_y3_severe.lag 1.44 0.0306 0.0319 45.0 0
4 less_90d_tr1.lag 1.28 0.0115 0.0119 108. 0
5 policonsumo2.lag 0.840 0.0185 0.0211 39.8 0 Code
i1b_df_out<-cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag",
"comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag"
), estimate = c(0.204663142678526, 0.00379136822637489, 1.43645954073343,
1.27844124422329, 0.840329282274339), std.error = c(0.0185085344588536,
0.0318597956273076, 0.0306100046937583, 0.011541793220903, 0.0185082077607319
), robust.se = c(0.020386616410505, 0.0335664648273184, 0.03190467043629,
0.0118920257355396, 0.0211189171816328), statistic = c(10.039093224566,
0.112951073217852, 45.0234878182453, 107.504076484012, 39.7903583335785
), p.value = c(1.02613077760236e-23, 0.910069344802066, 0, 0,
0))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i1$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.3904 1.0000 1.0000 0.9119 1.0000 1.0000 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.3896 1.0000 1.0000 0.9118 1.0000 1.0000
ploti1<-
ggplot( data= as.data.frame(i1$iiw.weight),aes(x= `i1$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
summary(i1b$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.0233 1.0000 1.0000 0.8543 1.0000 1.0000 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
#0.02329 1.00000 1.00000 0.85417 1.00000 1.00000
ploti1b<-
ggplot( data= as.data.frame(i1b$iiw.weight),aes(x= `i1b$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
i1c <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ #less_90d_tr.lag+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc_multtr,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2"),
maxfu= maxfu_multtr_df,#12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= rep(0,6),
formulanull= Surv(time.lag,time,event)~ 1,
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
i1d <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ #less_90d_tr.lag+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc_multtr,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2"),
maxfu= maxfu_multtr_df,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst=c(1,1,0,1,1,1),
formulanull= Surv(time.lag,time,event)~ 1,
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
i1c$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag, data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.30163 0.73961 0.01544 0.01298 -23.240
comp_bpsc_y2_moderate.lag -0.90364 0.40509 0.01902 0.01729 -52.268
comp_bpsc_y3_severe.lag -0.81206 0.44394 0.01987 0.01813 -44.786
less_90d_tr1.lag -0.11483 0.89152 0.01764 0.01581 -7.262
policonsumo2.lag -0.30996 0.73347 0.01596 0.01333 -23.251
p
tr_outcome.lag < 2e-16
comp_bpsc_y2_moderate.lag < 2e-16
comp_bpsc_y3_severe.lag < 2e-16
less_90d_tr1.lag 3.83e-13
policonsumo2.lag < 2e-16
Likelihood ratio test=7757 on 5 df, p=< 2.2e-16
n= 44272, number of events= 30959
(29 observations deleted due to missingness)Code
broom::tidy(i1c$m)# A tibble: 5 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag -0.302 0.0154 0.0130 -23.2 1.80e-119
2 comp_bpsc_y2_moderate.lag -0.904 0.0190 0.0173 -52.3 0
3 comp_bpsc_y3_severe.lag -0.812 0.0199 0.0181 -44.8 0
4 less_90d_tr1.lag -0.115 0.0176 0.0158 -7.26 3.83e- 13
5 policonsumo2.lag -0.310 0.0160 0.0133 -23.3 1.40e-119Code
i1c_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag",
"comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag"
), estimate = c(-0.302245037062739, -0.903956210807159, -0.813321897718714,
-0.114245961363377, -0.308855961637097), std.error = c(0.0154310846877583,
0.0190313654021939, 0.0198660721579724, 0.0176396124863469, 0.015967581763868
), robust.se = c(0.0129727086353161, 0.0173025650006909, 0.0181205187074852,
0.0158153935544036, 0.0133307124770294), statistic = c(-23.2985296717391,
-52.2440580787336, -44.8840295825939, -7.22371915503593, -23.1687512703688
), p.value = c(4.58804176323613e-120, 0, 0, 5.05846724994571e-13,
9.40892053149018e-119))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
i1d$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag, data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z p
tr_outcome.lag 0.09184 1.09619 0.01891 0.01472 6.240 4.36e-10
comp_bpsc_y2_moderate.lag -0.04619 0.95486 0.03189 0.02286 -2.021 0.0433
comp_bpsc_y3_severe.lag 0.39616 1.48611 0.03135 0.02234 17.736 < 2e-16
less_90d_tr1.lag 0.44917 1.56700 0.01418 0.01230 36.508 < 2e-16
policonsumo2.lag 0.23612 1.26632 0.01918 0.01518 15.553 < 2e-16
Likelihood ratio test=3294 on 5 df, p=< 2.2e-16
n= 43638, number of events= 30325
(663 observations deleted due to missingness)Code
broom::tidy(i1d$m)# A tibble: 5 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 0.0918 0.0189 0.0147 6.24 4.36e- 10
2 comp_bpsc_y2_moderate.lag -0.0462 0.0319 0.0229 -2.02 4.33e- 2
3 comp_bpsc_y3_severe.lag 0.396 0.0313 0.0223 17.7 2.23e- 70
4 less_90d_tr1.lag 0.449 0.0142 0.0123 36.5 8.35e-292
5 policonsumo2.lag 0.236 0.0192 0.0152 15.6 1.52e- 54Code
i1d_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag",
"comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag"
), estimate = c(0.0914594442348692, -0.0479067360824922, 0.395018880185181,
0.449421044097101, 0.237239909387438), std.error = c(0.0189119890640763,
0.0318880405199674, 0.0313463499911823, 0.0141743240908896, 0.0191755226262578
), robust.se = c(0.0147135065696911, 0.0228906898308705, 0.0223775814909527,
0.0123024239813926, 0.0151689053922935), statistic = c(6.21601953291473,
-2.09284807214001, 17.6524384614525, 36.5310970242003, 15.6398832514288
), p.value = c(5.09924589471951e-10, 0.0363627189586838, 9.74559709383605e-70,
3.55949649474628e-292, 3.89441388490098e-55))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i1c$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
1.000 1.000 3.046 2.660 4.152 5.104 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 1.000 1.000 3.051 2.660 4.155 5.100
ploti1c<-
ggplot( data= as.data.frame(i1c$iiw.weight),aes(x= `i1c$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
summary(i1d$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.3093 0.3093 0.3917 0.4846 0.6729 1.0473 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.3094 0.3094 0.3922 0.4848 0.6737 1.0491
ploti1d<-
ggplot( data= as.data.frame(i1d$iiw.weight),aes(x= `i1d$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
invisible("Now we use the 'data_mine_miss_proc' and try to dummify previous events")
i2 <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ treatment+#less_90d_tr.lag+
cluster(hash_key),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2", "treatment"),
maxfu= maxfu_df,#12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(0,0,0,0,0,0,0),
formulanull= Surv(time.lag,time,event)~ 1,
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
i2$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag + treatment, data = datacox, cluster = hash_key)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.20589 0.81392 0.01573 0.01912 -10.771
comp_bpsc_y2_moderate.lag -0.79538 0.45141 0.01931 0.02541 -31.300
comp_bpsc_y3_severe.lag -0.64342 0.52549 0.02049 0.02735 -23.522
less_90d_tr1.lag 0.16432 1.17859 0.01785 0.02055 7.997
policonsumo2.lag -0.49413 0.61010 0.01610 0.01925 -25.668
treatment -0.26432 0.76772 0.01253 0.01588 -16.644
p
tr_outcome.lag < 2e-16
comp_bpsc_y2_moderate.lag < 2e-16
comp_bpsc_y3_severe.lag < 2e-16
less_90d_tr1.lag 1.28e-15
policonsumo2.lag < 2e-16
treatment < 2e-16
Likelihood ratio test=18907 on 6 df, p=< 2.2e-16
n= 89925, number of events= 89925
(72548 observations deleted due to missingness)Code
broom::tidy(i2$m)# A tibble: 6 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag -0.206 0.0157 0.0191 -10.8 4.73e- 27
2 comp_bpsc_y2_moderate.lag -0.795 0.0193 0.0254 -31.3 4.67e-215
3 comp_bpsc_y3_severe.lag -0.643 0.0205 0.0274 -23.5 2.43e-122
4 less_90d_tr1.lag 0.164 0.0179 0.0205 8.00 1.28e- 15
5 policonsumo2.lag -0.494 0.0161 0.0193 -25.7 2.64e-145
6 treatment -0.264 0.0125 0.0159 -16.6 3.36e- 62Code
i2_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag",
"comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag",
"treatment"), estimate = c(-0.207821827365346, -0.795634053728106,
-0.645100128805644, 0.165468256440595, -0.492699340028085, -0.263268962510958
), std.error = c(0.0157193638381784, 0.0193245384394512, 0.0204926882142567,
0.0178456628682898, 0.0161016116581923, 0.0125485993898615),
robust.se = c(0.0190740908371374, 0.0254091936208088, 0.0272893127250465,
0.0205418969887765, 0.019247319556775, 0.0157928620459193
), statistic = c(-10.8955037039414, -31.3128415486796, -23.6392955478708,
8.05515948848355, -25.5983353201332, -16.6701236131537),
p.value = c(1.21094635824601e-27, 3.120318630146e-215, 1.52095150935808e-123,
7.93752420512666e-16, 1.5920600972545e-144, 2.1614709120083e-62
))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
invisible("Does really previous dropout protect dropout?")
invisible("maybe because lagfirst was 0")
invisible("Does really Severe previous biopsychosocial is a protective factor of dropout?")
invisible("Does really 0-90 days in previous treatment is a protective factor of dropout?")
invisible("Does really PSU in previous treatment is a protective factor of dropout?")
invisible("Does really more treatments protect dropout?")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i2$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
1.00 1.00 1.00 1.72 1.00 44.92 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 1.00 1.00 1.00 1.72 1.00 44.54
ploti2<-
ggplot( data= as.data.frame(i2$iiw.weight),aes(x= `i2$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
invisible("Some extreme weights")
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
i3 <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ policonsumo2+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2"),
maxfu= maxfu_df, #12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(0,0,0,0,0,0),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
i3$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag + policonsumo2, data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.283314 0.753283 0.014943 0.018002 -15.738
comp_bpsc_y2_moderate.lag -0.916036 0.400102 0.017794 0.022331 -41.021
comp_bpsc_y3_severe.lag -0.792492 0.452715 0.018677 0.023748 -33.370
less_90d_tr1.lag 0.116079 1.123085 0.017666 0.020221 5.741
policonsumo2.lag -0.572257 0.564251 0.015723 0.018761 -30.502
policonsumo2 0.164367 1.178647 0.007740 0.007572 21.706
p
tr_outcome.lag < 2e-16
comp_bpsc_y2_moderate.lag < 2e-16
comp_bpsc_y3_severe.lag < 2e-16
less_90d_tr1.lag 9.44e-09
policonsumo2.lag < 2e-16
policonsumo2 < 2e-16
Likelihood ratio test=18879 on 6 df, p=< 2.2e-16
n= 89925, number of events= 89925
(72548 observations deleted due to missingness)Code
broom::tidy(i3$m)# A tibble: 6 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag -0.283 0.0149 0.0180 -15.7 8.36e- 56
2 comp_bpsc_y2_moderate.lag -0.916 0.0178 0.0223 -41.0 0
3 comp_bpsc_y3_severe.lag -0.792 0.0187 0.0237 -33.4 3.68e-244
4 less_90d_tr1.lag 0.116 0.0177 0.0202 5.74 9.44e- 9
5 policonsumo2.lag -0.572 0.0157 0.0188 -30.5 2.42e-204
6 policonsumo2 0.164 0.00774 0.00757 21.7 1.80e-104Code
i3_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag", "comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag", "policonsumo2"), estimate = c(-0.285082046756843, -0.916585948362177,
-0.795077921451114, 0.11763808302626, -0.569791427138684, 0.164345823420033
), std.error = c(0.01492765131367, 0.0178053744947168, 0.0186611556090444,
0.0176615698524595, 0.0157351197601704, 0.00773975706029995),
robust.se = c(0.0179877493992649, 0.0223101496182252, 0.0236349676291152,
0.0202282450263194, 0.0186724156095066, 0.00757289335696934
), statistic = c(-15.8486779212352, -41.0838100168282, -33.639898895892,
5.81553579528025, -30.5151427139716, 21.7018536605676), p.value = c(1.43569566146437e-56,
0, 4.38118085650704e-248, 6.04399775632908e-09, 1.64100047755937e-204,
1.97056906465211e-104))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
invisible("Previous dropout prevents actual dropout?")
invisible("maybe because lagfirst was 0")
invisible("Severe biopsychosocial compromise is a risk factor of actual dropout?")
invisible("previous PSU prevents actual dropout?")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i3$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.7554 1.0000 1.0000 1.5860 1.0000 5.8803 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.7543 1.0000 1.0000 1.5863 1.0000 5.8794
ploti3<-
ggplot( data= as.data.frame(i3$iiw.weight),aes(x= `i3$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
invisible("Reasonable weights")
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
i3b <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ policonsumo2+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2"),
maxfu= maxfu_df,#12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(1,1,0,1,1,1),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
i3b$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag + policonsumo2, data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z p
tr_outcome.lag 0.171974 1.187647 0.018876 0.017987 9.561 < 2e-16
comp_bpsc_y2_moderate.lag 0.008456 1.008491 0.031855 0.029226 0.289 0.772
comp_bpsc_y3_severe.lag 0.597821 1.818153 0.030962 0.028960 20.643 < 2e-16
less_90d_tr1.lag 0.855059 2.351514 0.012903 0.014455 59.154 < 2e-16
policonsumo2.lag 0.123359 1.131291 0.019003 0.020580 5.994 2.04e-09
policonsumo2 0.062995 1.065021 0.007827 0.007610 8.277 < 2e-16
Likelihood ratio test=13404 on 6 df, p=< 2.2e-16
n= 87158, number of events= 87158
(75315 observations deleted due to missingness)Code
broom::tidy(i3b$m)# A tibble: 6 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 0.172 0.0189 0.0180 9.56 1.17e-21
2 comp_bpsc_y2_moderate.lag 0.00846 0.0319 0.0292 0.289 7.72e- 1
3 comp_bpsc_y3_severe.lag 0.598 0.0310 0.0290 20.6 1.13e-94
4 less_90d_tr1.lag 0.855 0.0129 0.0145 59.2 0
5 policonsumo2.lag 0.123 0.0190 0.0206 5.99 2.04e- 9
6 policonsumo2 0.0630 0.00783 0.00761 8.28 1.26e-16Code
i3b_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag", "comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag", "policonsumo2"), estimate = c(0.171524496903936, 0.00483620529764353, 0.596008396206245, 0.855542927694904, 0.125702439454889, 0.0627347379496959), std.error = c(0.018874296525738, 0.0318585598048639, 0.0309630894506515, 0.0128979749956164, 0.0189962053653323, 0.00782722711322035), robust.se = c(0.0179913060388931, 0.0292985300929012, 0.0290119903127091, 0.0144451828760062, 0.020549409265887, 0.00759613906581795), statistic = c(9.53374349439331, 0.165066482253842, 20.543519757938, 59.2268671873984, 6.11708287223425, 8.2587663819897), p.value = c(1.51711569984396e-21, 0.868891654364147, 8.79518334988573e-94, 0, 9.53038111046381e-10, 1.47185680747056e-16))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
invisible("previous dropout transform to nothing; 2024-04-16: dropping out previously is a risk factor")
invisible("previous PSU still protects an actual dropout; 2024-04-16: previous PSU is a risk factor")
invisible("previous severe biopsuchosocial still is a risk factor; 2024-04-16: severe biopsychosocial, risk factor (80%+)")
invisible("now less than 90 days in treatment its a risk factor; 2024-04-16: less than 90 days tr. previously, risk factor of being observed posteriorly (readmission) (x2.3)")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i3b$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.1635 1.0000 1.0000 0.9053 1.0000 1.0000 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.1634 1.0000 1.0000 0.9055 1.0000 1.0000
ploti3b<-
ggplot( data= as.data.frame(i3b$iiw.weight),aes(x= `i3b$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
invisible("Many 0s or near 0s")
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
i3_nolag <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ policonsumo2+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2"),
maxfu= maxfu_df, #12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
#lagfirst= rep(1,6),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)Error in shift(.SD, fill = lagfirst[col]): entrada en evaluacion: recursivo por defecto o problemas anteriores?
Code
invisible("Here, i didnt specified lagfirst")
# Error in shift(.SD, fill = lagfirst[col]) :
# entrada en evaluacion: recursivo por defecto o problemas anteriores?
invisible("It gives an error")
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
invisible("We format the data to include non-events; 2024-04-16: this is no longer useful because we are including every record (all event=1)")
# data_mine2_miss <- Base_fiscalia_v15f_grant_23_24_long2_miss_proc
#
# data_mine2_miss_proc <-
# data_mine2_miss %>%
# dplyr::mutate(comp_bpsc_y2_moderate= ifelse(as.character(comp_bpsc_y)=="2-Moderate",1,0),
# comp_bpsc_y3_severe= ifelse(as.character(comp_bpsc_y)=="3-Severe",1,0),
# less_90d_tr1= ifelse(less_90d_tr==1,1,0)) %>% data.frame()
i4 <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ treatment.lag+
cluster(hash_key),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2", "treatment"),
maxfu= maxfu_df,#12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= rep(0,7),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
i4$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag + treatment.lag, data = datacox, cluster = hash_key)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.20589 0.81392 0.01573 0.01912 -10.771
comp_bpsc_y2_moderate.lag -0.79538 0.45141 0.01931 0.02541 -31.300
comp_bpsc_y3_severe.lag -0.64342 0.52549 0.02049 0.02735 -23.522
less_90d_tr1.lag 0.16432 1.17859 0.01785 0.02055 7.997
policonsumo2.lag -0.49413 0.61010 0.01610 0.01925 -25.668
treatment.lag -0.26432 0.76772 0.01253 0.01588 -16.644
p
tr_outcome.lag < 2e-16
comp_bpsc_y2_moderate.lag < 2e-16
comp_bpsc_y3_severe.lag < 2e-16
less_90d_tr1.lag 1.28e-15
policonsumo2.lag < 2e-16
treatment.lag < 2e-16
Likelihood ratio test=18907 on 6 df, p=< 2.2e-16
n= 89925, number of events= 89925
(72548 observations deleted due to missingness)Code
broom::tidy(i4$m)# A tibble: 6 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag -0.206 0.0157 0.0191 -10.8 4.73e- 27
2 comp_bpsc_y2_moderate.lag -0.795 0.0193 0.0254 -31.3 4.67e-215
3 comp_bpsc_y3_severe.lag -0.643 0.0205 0.0274 -23.5 2.43e-122
4 less_90d_tr1.lag 0.164 0.0179 0.0205 8.00 1.28e- 15
5 policonsumo2.lag -0.494 0.0161 0.0193 -25.7 2.64e-145
6 treatment.lag -0.264 0.0125 0.0159 -16.6 3.36e- 62Code
i4_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag",
"comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag",
"treatment.lag"), estimate = c(-0.207821827365346, -0.795634053728106,
-0.645100128805644, 0.165468256440595, -0.492699340028085, -0.263268962510958
), std.error = c(0.0157193638381784, 0.0193245384394512, 0.0204926882142567,
0.0178456628682898, 0.0161016116581923, 0.0125485993898615),
robust.se = c(0.0190740908371374, 0.0254091936208087, 0.0272893127250465,
0.0205418969887765, 0.019247319556775, 0.0157928620459193
), statistic = c(-10.8955037039414, -31.3128415486796, -23.6392955478708,
8.05515948848355, -25.5983353201332, -16.6701236131537),
p.value = c(1.21094635824605e-27, 3.12031863014138e-215,
1.52095150935808e-123, 7.93752420512701e-16, 1.59206009725477e-144,
2.16147091200972e-62))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
invisible("Previous dropout protects posterior dropout, but 95% CI does not replicate that; prev. dropouts protects posterior obs.")
invisible("Moderate previous biopsychosocial status protects actual dropout; both severe and moderate pro")
invisible("Severe previous biopsychosocial status protects actual dropout")
invisible("0-90 days in previous tr. predicts actual dropout")
invisible("previous PSU protects actual dropout")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i4$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
1.00 1.00 1.00 1.72 1.00 44.92 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 1.00 1.00 1.00 1.72 1.00 44.54 #2024-04-16: muy bizarro que haya un caso con probabilidades muy altas
ploti4<-
ggplot( data= as.data.frame(i4$iiw.weight),aes(x= `i4$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
invisible("Reasonable, some positive outliers but very big")
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
i4b <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ treatment.lag+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2", "treatment"),
maxfu= maxfu_df,#12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(1,1,0,1,1,1,1),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
i4b$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag + treatment.lag, data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.203911 1.226189 0.018495 0.020208 10.091
comp_bpsc_y2_moderate.lag -0.001234 0.998767 0.031861 0.033417 -0.037
comp_bpsc_y3_severe.lag 1.426373 4.163571 0.030614 0.031776 44.888
less_90d_tr1.lag 1.273223 3.572347 0.011538 0.011872 107.246
policonsumo2.lag 0.846049 2.330422 0.018511 0.020888 40.503
treatment.lag 0.145624 1.156762 0.011528 0.015317 9.507
p
tr_outcome.lag <2e-16
comp_bpsc_y2_moderate.lag 0.971
comp_bpsc_y3_severe.lag <2e-16
less_90d_tr1.lag <2e-16
policonsumo2.lag <2e-16
treatment.lag <2e-16
Likelihood ratio test=61967 on 6 df, p=< 2.2e-16
n= 159556, number of events= 87158
(2917 observations deleted due to missingness)Code
broom::tidy(i4b$m)# A tibble: 6 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 0.204 0.0185 0.0202 10.1 6.07e-24
2 comp_bpsc_y2_moderate.lag -0.00123 0.0319 0.0334 -0.0369 9.71e- 1
3 comp_bpsc_y3_severe.lag 1.43 0.0306 0.0318 44.9 0
4 less_90d_tr1.lag 1.27 0.0115 0.0119 107. 0
5 policonsumo2.lag 0.846 0.0185 0.0209 40.5 0
6 treatment.lag 0.146 0.0115 0.0153 9.51 1.96e-21Code
i4b_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag",
"comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag",
"treatment.lag"), estimate = c(0.204102531211081, -0.000387761814109535,
1.42748577725764, 1.27218207100866, 0.846223498441861, 0.145344423972201
), std.error = c(0.0184955717163036, 0.0318615324506995, 0.0306160739199961,
0.011538166117299, 0.0185111088069143, 0.0115312169046261), robust.se = c(0.0202100857453012,
0.0334267319824861, 0.0317905279138096, 0.0118704233236429, 0.0208804874610975,
0.0153330805627965), statistic = c(10.0990433085389, -0.0116003507107037,
44.9028648133129, 107.172426485818, 40.5269991909175, 9.47914043606206
), p.value = c(5.5783718456834e-24, 0.990744466851805, 0, 0,
0, 2.56386267293478e-21))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
invisible("Previous dropout now predicts actual dropout")
invisible("Moderate previous biopsychosocial status does not [2024-04-16] predict actual dropout")
invisible("Severe previous biopsychosocial status predict actual dropout")
invisible("0-90 days in previous tr. predicts actual dropout")
invisible("previous PSU predicts posterior dropout")
invisible("more treatments can be a protective factor for subsequent treatments [2024-04-16]")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i4b$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.008675 1.000000 1.000000 0.853592 1.000000 1.023063 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.008687 1.000000 1.000000 0.853483 1.000000 1.022155
ploti4b<-
ggplot( data= as.data.frame(i4b$iiw.weight),aes(x= `i4b$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
invisible("Reasonable")
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#2024-04-16: no longer useful
#
# data_mine2_miss_proc_multtr<-
# data_mine2_miss_proc %>%
# dplyr::group_by(hash_key) %>% dplyr::mutate(n=n()) %>% dplyr::ungroup() %>% dplyr::filter(case_when(n>1~ T,T~F)) %>% data.frame()
#
#
# paste0("The original database had ",
# format(length(unique(data_mine2_miss_proc$hash_key)), big.mark=","),
# " patients and ",
# format(length(data_mine2_miss_proc$hash_key), big.mark=","),
# " observations with only events, but keeping patients with more than one treatment left only ",
# format(length(unique(data_mine2_miss_proc_multtr$hash_key)), big.mark=","),
# " patients and ",
# format(length(data_mine2_miss_proc_multtr$hash_key), big.mark=","),
# " observations"
# )
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
i4c <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ treatment.lag+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc_multtr,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2", "treatment"),
maxfu= maxfu_multtr_df,#12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= rep(0,7),# c(1,1,0,1,1,1,1),
#formulanull= Surv(time.lag,time,event)~ 1,
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
invisible("FORMULANULL entrega ponderaciones y coeficientes iguales a sin formulanull")
i4c$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag + treatment.lag, data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.02895 1.02937 0.01743 0.01596 1.814
comp_bpsc_y2_moderate.lag -0.28716 0.75039 0.02330 0.02407 -11.928
comp_bpsc_y3_severe.lag -0.12854 0.87938 0.02427 0.02515 -5.112
less_90d_tr1.lag 0.12332 1.13125 0.01801 0.01722 7.162
policonsumo2.lag -0.18087 0.83454 0.01788 0.01620 -11.167
treatment.lag -1.33538 0.26306 0.01551 0.02788 -47.895
p
tr_outcome.lag 0.0697
comp_bpsc_y2_moderate.lag < 2e-16
comp_bpsc_y3_severe.lag 3.19e-07
less_90d_tr1.lag 7.94e-13
policonsumo2.lag < 2e-16
treatment.lag < 2e-16
Likelihood ratio test=16880 on 6 df, p=< 2.2e-16
n= 44272, number of events= 30959
(29 observations deleted due to missingness)Code
broom::tidy(i4c$m)# A tibble: 6 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 0.0290 0.0174 0.0160 1.81 6.97e- 2
2 comp_bpsc_y2_moderate.lag -0.287 0.0233 0.0241 -11.9 8.48e-33
3 comp_bpsc_y3_severe.lag -0.129 0.0243 0.0251 -5.11 3.19e- 7
4 less_90d_tr1.lag 0.123 0.0180 0.0172 7.16 7.94e-13
5 policonsumo2.lag -0.181 0.0179 0.0162 -11.2 5.91e-29
6 treatment.lag -1.34 0.0155 0.0279 -47.9 0 Code
i4c_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag",
"comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag",
"treatment.lag"), estimate = c(0.0283175141261873, -0.286052899938312,
-0.12762290937749, 0.123676646115759, -0.18046907845805, -1.33577820010306
), std.error = c(0.0174225204858456, 0.0232998089273985, 0.0242539101271994,
0.0180040024564168, 0.0178722433414989, 0.0155176207475381),
robust.se = c(0.0159600395125907, 0.0240699432379503, 0.0251576808692291,
0.0172158438569811, 0.016189953421006, 0.0278980099478121
), statistic = c(1.77427594109952, -11.884236581302, -5.07292027595389,
7.18388521313221, -11.1469794733255, -47.8807700836676),
p.value = c(0.0760175181238697, 1.42935024171782e-32, 3.91756657551751e-07,
6.77577589606902e-13, 7.40784798532889e-29, 0))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
invisible("Risk factors: previous dropout, <90 days previous treatment; protective: previous PSU!!!!, previous treatments")
i4d <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ treatment.lag+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc_multtr,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2", "treatment"),
maxfu= maxfu_multtr_df,#12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(1,1,0,1,1,1,1),#
#formulanull= Surv(time.lag,time,event)~ 1,
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
i4d$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag + treatment.lag, data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.132783 1.142003 0.019040 0.016285 8.154
comp_bpsc_y2_moderate.lag 0.003703 1.003710 0.031880 0.025639 0.144
comp_bpsc_y3_severe.lag 0.479705 1.615598 0.031391 0.025947 18.488
less_90d_tr1.lag 0.589226 1.802592 0.014530 0.014917 39.500
policonsumo2.lag 0.103722 1.109292 0.019252 0.018027 5.754
treatment.lag -1.346903 0.260044 0.016631 0.032231 -41.789
p
tr_outcome.lag 3.52e-16
comp_bpsc_y2_moderate.lag 0.885
comp_bpsc_y3_severe.lag < 2e-16
less_90d_tr1.lag < 2e-16
policonsumo2.lag 8.72e-09
treatment.lag < 2e-16
Likelihood ratio test=11563 on 6 df, p=< 2.2e-16
n= 43638, number of events= 30325
(663 observations deleted due to missingness)Code
broom::tidy(i4d$m)# A tibble: 6 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 0.133 0.0190 0.0163 8.15 3.52e-16
2 comp_bpsc_y2_moderate.lag 0.00370 0.0319 0.0256 0.144 8.85e- 1
3 comp_bpsc_y3_severe.lag 0.480 0.0314 0.0259 18.5 2.58e-76
4 less_90d_tr1.lag 0.589 0.0145 0.0149 39.5 0
5 policonsumo2.lag 0.104 0.0193 0.0180 5.75 8.72e- 9
6 treatment.lag -1.35 0.0166 0.0322 -41.8 0 Code
i4d_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag",
"comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag",
"treatment.lag"), estimate = c(0.132485185827427, 0.00411115497188839,
0.481109354943001, 0.58947698342646, 0.105500181664414, -1.34747522983601
), std.error = c(0.0190393190327229, 0.0318822421857093, 0.0313896070698578,
0.0145276202543287, 0.0192443206789934, 0.0166349222671111),
robust.se = c(0.0162768310200937, 0.0256183384356621, 0.0259237137593143,
0.0149119908677786, 0.0180095159064235, 0.0322392535580045
), statistic = c(8.13949506902635, 0.160477034145409, 18.5586586632535,
39.5304013161775, 5.85802429185703, -41.7961050931792), p.value = c(3.96929765929241e-16,
0.87250531125718, 6.94162601178333e-77, 0, 4.68406148676732e-09,
0))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
invisible("Razonable: abandono previo, compromiso severo previo, menos de 90 días y policonsumo previo son predictores de readmisiones; no así el recuento de tratamientos (esto con lag 1's)")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i4c$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.19 0.19 0.86 4.91 1.14 49523.91 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.19 0.19 0.86 4.91 1.13 49631.97
ploti4c<-
ggplot( data= as.data.frame(i4c$iiw.weight),aes(x= `i4c$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
invisible("highly extreme values")
summary(i4d$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.131 0.131 0.210 1.515 0.379 18136.282 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.130 0.130 0.210 1.516 0.378 18179.444
ploti4d<-
ggplot( data= as.data.frame(i4d$iiw.weight),aes(x= `i4d$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
invisible(" highly extreme values")
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
i5 <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ policonsumo2+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2"),
maxfu= maxfu_df,#12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= rep(0,6),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
i5$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag + policonsumo2, data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.283314 0.753283 0.014943 0.018002 -15.738
comp_bpsc_y2_moderate.lag -0.916036 0.400102 0.017794 0.022331 -41.021
comp_bpsc_y3_severe.lag -0.792492 0.452715 0.018677 0.023748 -33.370
less_90d_tr1.lag 0.116079 1.123085 0.017666 0.020221 5.741
policonsumo2.lag -0.572257 0.564251 0.015723 0.018761 -30.502
policonsumo2 0.164367 1.178647 0.007740 0.007572 21.706
p
tr_outcome.lag < 2e-16
comp_bpsc_y2_moderate.lag < 2e-16
comp_bpsc_y3_severe.lag < 2e-16
less_90d_tr1.lag 9.44e-09
policonsumo2.lag < 2e-16
policonsumo2 < 2e-16
Likelihood ratio test=18879 on 6 df, p=< 2.2e-16
n= 89925, number of events= 89925
(72548 observations deleted due to missingness)Code
broom::tidy(i5$m)# A tibble: 6 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag -0.283 0.0149 0.0180 -15.7 8.36e- 56
2 comp_bpsc_y2_moderate.lag -0.916 0.0178 0.0223 -41.0 0
3 comp_bpsc_y3_severe.lag -0.792 0.0187 0.0237 -33.4 3.68e-244
4 less_90d_tr1.lag 0.116 0.0177 0.0202 5.74 9.44e- 9
5 policonsumo2.lag -0.572 0.0157 0.0188 -30.5 2.42e-204
6 policonsumo2 0.164 0.00774 0.00757 21.7 1.80e-104Code
i5_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag",
"comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag",
"policonsumo2"), estimate = c(-0.285082046756843, -0.916585948362177,
-0.795077921451114, 0.11763808302626, -0.569791427138684, 0.164345823420033
), std.error = c(0.01492765131367, 0.0178053744947168, 0.0186611556090444,
0.0176615698524595, 0.0157351197601704, 0.00773975706029995),
robust.se = c(0.0179877493992649, 0.0223101496182252, 0.0236349676291152,
0.0202282450263194, 0.0186724156095066, 0.00757289335696934
), statistic = c(-15.8486779212352, -41.0838100168282, -33.639898895892,
5.81553579528025, -30.5151427139716, 21.7018536605676), p.value = c(1.43569566146437e-56,
0, 4.38118085650704e-248, 6.04399775632908e-09, 1.64100047755937e-204,
1.97056906465211e-104))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
invisible("Previous dropout protects posterior dropout, but 95% CI does not replicate that; 2024-04-16= protects from being observed and 95% cis too")
invisible("Moderate previous biopsychosocial status protects actual dropout")
invisible("Severe previous biopsychosocial status protects actual dropout")
invisible("0-90 days in previous tr. predicts actual dropout")
invisible("previous PSU protects actual dropout")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i5$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.7554 1.0000 1.0000 1.5860 1.0000 5.8803 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.7543 1.0000 1.0000 1.5863 1.0000 5.8794
ploti5<-
ggplot( data= as.data.frame(i5$iiw.weight),aes(x= `i5$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
invisible("Reasonable, but still there are some bit extreme values")
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
i5b <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y2_moderate.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ policonsumo2.lag+ policonsumo2+
cluster(id),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc,
invariant= "id",
lagvars= c("time", "tr_outcome", "comp_bpsc_y2_moderate", "comp_bpsc_y3_severe", "less_90d_tr1","policonsumo2"),
maxfu= maxfu_df,# 12*12,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(1,1,0,1,1,1,1),
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
i5b$mCall:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
comp_bpsc_y2_moderate.lag + comp_bpsc_y3_severe.lag + less_90d_tr1.lag +
policonsumo2.lag + policonsumo2, data = datacox, cluster = id)
coef exp(coef) se(coef) robust se z p
tr_outcome.lag 0.171974 1.187647 0.018876 0.017987 9.561 < 2e-16
comp_bpsc_y2_moderate.lag 0.008456 1.008491 0.031855 0.029226 0.289 0.772
comp_bpsc_y3_severe.lag 0.597821 1.818153 0.030962 0.028960 20.643 < 2e-16
less_90d_tr1.lag 0.855059 2.351514 0.012903 0.014455 59.154 < 2e-16
policonsumo2.lag 0.123359 1.131291 0.019003 0.020580 5.994 2.04e-09
policonsumo2 0.062995 1.065021 0.007827 0.007610 8.277 < 2e-16
Likelihood ratio test=13404 on 6 df, p=< 2.2e-16
n= 87158, number of events= 87158
(75315 observations deleted due to missingness)Code
broom::tidy(i5b$m)# A tibble: 6 x 6
term estimate std.error robust.se statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 0.172 0.0189 0.0180 9.56 1.17e-21
2 comp_bpsc_y2_moderate.lag 0.00846 0.0319 0.0292 0.289 7.72e- 1
3 comp_bpsc_y3_severe.lag 0.598 0.0310 0.0290 20.6 1.13e-94
4 less_90d_tr1.lag 0.855 0.0129 0.0145 59.2 0
5 policonsumo2.lag 0.123 0.0190 0.0206 5.99 2.04e- 9
6 policonsumo2 0.0630 0.00783 0.00761 8.28 1.26e-16Code
i5b_df_out<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y2_moderate.lag",
"comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag",
"policonsumo2"), estimate = c(0.171524496903936, 0.00483620529764353,
0.596008396206245, 0.855542927694904, 0.125702439454889, 0.0627347379496959
), std.error = c(0.018874296525738, 0.0318585598048639, 0.0309630894506515,
0.0128979749956164, 0.0189962053653323, 0.00782722711322035),
robust.se = c(0.0179913060388931, 0.0292985300929012, 0.0290119903127091,
0.0144451828760062, 0.020549409265887, 0.00759613906581795
), statistic = c(9.53374349439331, 0.165066482253842, 20.543519757938,
59.2268671873984, 6.11708287223425, 8.2587663819897), p.value = c(1.51711569984396e-21,
0.868891654364147, 8.79518334988573e-94, 0, 9.53038111046381e-10,
1.47185680747056e-16))%>%
dplyr::mutate(lo_95ci= estimate- 1.96 * robust.se) %>%
dplyr::mutate(up_95ci= estimate+ 1.96 * robust.se) %>%
dplyr::mutate(HR= estimate) %>%
dplyr::mutate_at(c("HR","lo_95ci","up_95ci"),~exp(.))
invisible("Previous dropout now predicts actual dropout")
invisible("Moderate previous biopsychosocial status predict actual dropout")
invisible("Severe previous biopsychosocial status predict actual dropout")
invisible("0-90 days in previous tr. predicts actual dropout")
invisible("previous PSU protects actual dropout")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
summary(i5b$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.1635 1.0000 1.0000 0.9053 1.0000 1.0000 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.1634 1.0000 1.0000 0.9055 1.0000 1.0000
invisible("Many low values")
ploti5b<-
ggplot( data= as.data.frame(i5b$iiw.weight),aes(x= `i5b$iiw.weight`))+
geom_histogram(bins=60)+
theme_classic()+ xlim(0,10)
invisible("Reasonable, but still there are some bit extreme values")Code
iiw_coefs_df<-
dplyr::bind_rows(
cbind.data.frame(w= "i0", i0_df_out),
cbind.data.frame(w= "i0b", i0b_df_out),
cbind.data.frame(w= "i0c", i0c_df_out),
cbind.data.frame(w= "i0d", i0d_df_out),
cbind.data.frame(w= "i1", i1_df_out),
cbind.data.frame(w= "i1b", i1b_df_out),
cbind.data.frame(w= "i1c", i1c_df_out),
cbind.data.frame(w= "i1d", i1d_df_out),
cbind.data.frame(w= "i2", i2_df_out),
cbind.data.frame(w= "i3", i3_df_out),
cbind.data.frame(w= "i3b", i3b_df_out),
cbind.data.frame(w= "i4", i4_df_out),
cbind.data.frame(w= "i4b", i4b_df_out),
cbind.data.frame(w= "i4c", i4c_df_out),
cbind.data.frame(w= "i4d", i4d_df_out),
cbind.data.frame(w= "i5", i5_df_out),
cbind.data.frame(w= "i5b", i5b_df_out)
) %>% dplyr::select(w, term, HR, lo_95ci, up_95ci, p.value)
iiw_coefs_df %>%
dplyr::mutate_at(c("HR", "lo_95ci", "up_95ci"), ~ sprintf("%1.2f",.)) %>%
dplyr::mutate_at(c("p.value"), ~ ifelse(.<.001,"<0.001",sprintf("%1.3f",.))) %>%
{
write.table(., file= paste0(getwd(),"/_proposal_grant/2023/iiws_coefs_24_04_18.csv"), dec=",", sep="\t")
knitr::kable(., size=10, format="markdown",caption= "Different weights, coefficients", escape=T)
}| w | term | HR | lo_95ci | up_95ci | p.value |
|---|---|---|---|---|---|
| i0 | tr_outcome.lag | 0.15 | 0.15 | 0.16 | <0.001 |
| i0b | tr_outcome.lag | 2.91 | 2.82 | 3.01 | <0.001 |
| i0c | tr_outcome.lag | 0.47 | 0.47 | 0.48 | <0.001 |
| i0d | tr_outcome.lag | 1.32 | 1.28 | 1.35 | <0.001 |
| i1 | tr_outcome.lag | 1.20 | 1.15 | 1.24 | <0.001 |
| i1 | comp_bpsc_y2_moderate.lag | 1.17 | 1.10 | 1.24 | <0.001 |
| i1 | comp_bpsc_y3_severe.lag | 1.69 | 1.58 | 1.80 | <0.001 |
| i1 | less_90d_tr1.lag | 1.04 | 1.00 | 1.08 | 0.032 |
| i1 | policonsumo2.lag | 1.22 | 1.17 | 1.27 | <0.001 |
| i1b | tr_outcome.lag | 1.23 | 1.18 | 1.28 | <0.001 |
| i1b | comp_bpsc_y2_moderate.lag | 1.00 | 0.94 | 1.07 | 0.910 |
| i1b | comp_bpsc_y3_severe.lag | 4.21 | 3.95 | 4.48 | <0.001 |
| i1b | less_90d_tr1.lag | 3.59 | 3.51 | 3.68 | <0.001 |
| i1b | policonsumo2.lag | 2.32 | 2.22 | 2.42 | <0.001 |
| i1c | tr_outcome.lag | 0.74 | 0.72 | 0.76 | <0.001 |
| i1c | comp_bpsc_y2_moderate.lag | 0.40 | 0.39 | 0.42 | <0.001 |
| i1c | comp_bpsc_y3_severe.lag | 0.44 | 0.43 | 0.46 | <0.001 |
| i1c | less_90d_tr1.lag | 0.89 | 0.86 | 0.92 | <0.001 |
| i1c | policonsumo2.lag | 0.73 | 0.72 | 0.75 | <0.001 |
| i1d | tr_outcome.lag | 1.10 | 1.06 | 1.13 | <0.001 |
| i1d | comp_bpsc_y2_moderate.lag | 0.95 | 0.91 | 1.00 | 0.036 |
| i1d | comp_bpsc_y3_severe.lag | 1.48 | 1.42 | 1.55 | <0.001 |
| i1d | less_90d_tr1.lag | 1.57 | 1.53 | 1.61 | <0.001 |
| i1d | policonsumo2.lag | 1.27 | 1.23 | 1.31 | <0.001 |
| i2 | tr_outcome.lag | 0.81 | 0.78 | 0.84 | <0.001 |
| i2 | comp_bpsc_y2_moderate.lag | 0.45 | 0.43 | 0.47 | <0.001 |
| i2 | comp_bpsc_y3_severe.lag | 0.52 | 0.50 | 0.55 | <0.001 |
| i2 | less_90d_tr1.lag | 1.18 | 1.13 | 1.23 | <0.001 |
| i2 | policonsumo2.lag | 0.61 | 0.59 | 0.63 | <0.001 |
| i2 | treatment | 0.77 | 0.75 | 0.79 | <0.001 |
| i3 | tr_outcome.lag | 0.75 | 0.73 | 0.78 | <0.001 |
| i3 | comp_bpsc_y2_moderate.lag | 0.40 | 0.38 | 0.42 | <0.001 |
| i3 | comp_bpsc_y3_severe.lag | 0.45 | 0.43 | 0.47 | <0.001 |
| i3 | less_90d_tr1.lag | 1.12 | 1.08 | 1.17 | <0.001 |
| i3 | policonsumo2.lag | 0.57 | 0.55 | 0.59 | <0.001 |
| i3 | policonsumo2 | 1.18 | 1.16 | 1.20 | <0.001 |
| i3b | tr_outcome.lag | 1.19 | 1.15 | 1.23 | <0.001 |
| i3b | comp_bpsc_y2_moderate.lag | 1.00 | 0.95 | 1.06 | 0.869 |
| i3b | comp_bpsc_y3_severe.lag | 1.81 | 1.71 | 1.92 | <0.001 |
| i3b | less_90d_tr1.lag | 2.35 | 2.29 | 2.42 | <0.001 |
| i3b | policonsumo2.lag | 1.13 | 1.09 | 1.18 | <0.001 |
| i3b | policonsumo2 | 1.06 | 1.05 | 1.08 | <0.001 |
| i4 | tr_outcome.lag | 0.81 | 0.78 | 0.84 | <0.001 |
| i4 | comp_bpsc_y2_moderate.lag | 0.45 | 0.43 | 0.47 | <0.001 |
| i4 | comp_bpsc_y3_severe.lag | 0.52 | 0.50 | 0.55 | <0.001 |
| i4 | less_90d_tr1.lag | 1.18 | 1.13 | 1.23 | <0.001 |
| i4 | policonsumo2.lag | 0.61 | 0.59 | 0.63 | <0.001 |
| i4 | treatment.lag | 0.77 | 0.75 | 0.79 | <0.001 |
| i4b | tr_outcome.lag | 1.23 | 1.18 | 1.28 | <0.001 |
| i4b | comp_bpsc_y2_moderate.lag | 1.00 | 0.94 | 1.07 | 0.991 |
| i4b | comp_bpsc_y3_severe.lag | 4.17 | 3.92 | 4.44 | <0.001 |
| i4b | less_90d_tr1.lag | 3.57 | 3.49 | 3.65 | <0.001 |
| i4b | policonsumo2.lag | 2.33 | 2.24 | 2.43 | <0.001 |
| i4b | treatment.lag | 1.16 | 1.12 | 1.19 | <0.001 |
| i4c | tr_outcome.lag | 1.03 | 1.00 | 1.06 | 0.076 |
| i4c | comp_bpsc_y2_moderate.lag | 0.75 | 0.72 | 0.79 | <0.001 |
| i4c | comp_bpsc_y3_severe.lag | 0.88 | 0.84 | 0.92 | <0.001 |
| i4c | less_90d_tr1.lag | 1.13 | 1.09 | 1.17 | <0.001 |
| i4c | policonsumo2.lag | 0.83 | 0.81 | 0.86 | <0.001 |
| i4c | treatment.lag | 0.26 | 0.25 | 0.28 | <0.001 |
| i4d | tr_outcome.lag | 1.14 | 1.11 | 1.18 | <0.001 |
| i4d | comp_bpsc_y2_moderate.lag | 1.00 | 0.95 | 1.06 | 0.873 |
| i4d | comp_bpsc_y3_severe.lag | 1.62 | 1.54 | 1.70 | <0.001 |
| i4d | less_90d_tr1.lag | 1.80 | 1.75 | 1.86 | <0.001 |
| i4d | policonsumo2.lag | 1.11 | 1.07 | 1.15 | <0.001 |
| i4d | treatment.lag | 0.26 | 0.24 | 0.28 | <0.001 |
| i5 | tr_outcome.lag | 0.75 | 0.73 | 0.78 | <0.001 |
| i5 | comp_bpsc_y2_moderate.lag | 0.40 | 0.38 | 0.42 | <0.001 |
| i5 | comp_bpsc_y3_severe.lag | 0.45 | 0.43 | 0.47 | <0.001 |
| i5 | less_90d_tr1.lag | 1.12 | 1.08 | 1.17 | <0.001 |
| i5 | policonsumo2.lag | 0.57 | 0.55 | 0.59 | <0.001 |
| i5 | policonsumo2 | 1.18 | 1.16 | 1.20 | <0.001 |
| i5b | tr_outcome.lag | 1.19 | 1.15 | 1.23 | <0.001 |
| i5b | comp_bpsc_y2_moderate.lag | 1.00 | 0.95 | 1.06 | 0.869 |
| i5b | comp_bpsc_y3_severe.lag | 1.81 | 1.71 | 1.92 | <0.001 |
| i5b | less_90d_tr1.lag | 2.35 | 2.29 | 2.42 | <0.001 |
| i5b | policonsumo2.lag | 1.13 | 1.09 | 1.18 | <0.001 |
| i5b | policonsumo2 | 1.06 | 1.05 | 1.08 | <0.001 |
Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
iiw_df<-
dplyr::bind_rows(
cbind.data.frame(w= "i0", iiw=i0$iiw.weight),
cbind.data.frame(w= "i0b", iiw=i0b$iiw.weight),
cbind.data.frame(w= "i0c", iiw=i0c$iiw.weight),
cbind.data.frame(w= "i0d", iiw=i0d$iiw.weight),
cbind.data.frame(w= "i1", iiw=i1$iiw.weight),
cbind.data.frame(w= "i1b", iiw=i1b$iiw.weight),
cbind.data.frame(w= "i1c", iiw=i1c$iiw.weight),
cbind.data.frame(w= "i1d", iiw=i1d$iiw.weight),
cbind.data.frame(w= "i2", iiw=i2$iiw.weight),
cbind.data.frame(w= "i3", iiw=i3$iiw.weight),
cbind.data.frame(w= "i3b", iiw=i3b$iiw.weight),
cbind.data.frame(w= "i4", iiw=i4$iiw.weight),
cbind.data.frame(w= "i4b", iiw=i4b$iiw.weight),
cbind.data.frame(w= "i4c", iiw=i4c$iiw.weight),
cbind.data.frame(w= "i4d", iiw=i4d$iiw.weight),
cbind.data.frame(w= "i5", iiw=i5$iiw.weight),
cbind.data.frame(w= "i5b", iiw=i5b$iiw.weight)
)
psych::describeBy(iiw~w, data=iiw_df, mat=T) %>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/iiws_24_04_18.csv"), dec=",", sep="\t")
knitr::kable(., size=10, format="markdown",caption= "Different weights, summary descriptives", escape=T)
}| item | group1 | vars | n | mean | sd | median | trimmed | mad | min | max | range | skew | kurtosis | se | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| iiw1 | 1 | i0 | 1 | 90075 | 1.8464076 | 1.9802604 | 1.0000000 | 1.3731192 | 0.0000000 | 1.0000000 | 6.479385 | 5.479385e+00 | 1.9121364 | 1.6562841 | 0.0065981 |
| iiw2 | 2 | i0b | 1 | 90075 | 0.8986002 | 0.2372356 | 1.0000000 | 0.9553002 | 0.0000000 | 0.3435686 | 1.000000 | 6.564314e-01 | -1.9121364 | 1.6562841 | 0.0007905 |
| iiw3 | 3 | i0c | 1 | 30988 | 1.4988009 | 0.5526275 | 1.0000000 | 1.4846252 | 0.0000000 | 1.0000000 | 2.111044 | 1.111044e+00 | 0.2052710 | -1.9579270 | 0.0031393 |
| iiw4 | 4 | i0d | 1 | 30988 | 0.7889996 | 0.0784465 | 0.7598355 | 0.7662774 | 0.0000000 | 0.7598355 | 1.000000 | 2.401645e-01 | 2.3178907 | 3.3727262 | 0.0004456 |
| iiw5 | 5 | i1 | 1 | 90075 | 0.9119083 | 0.1872022 | 1.0000000 | 0.9587877 | 0.0000000 | 0.3903843 | 1.000000 | 6.096157e-01 | -1.8254047 | 1.6834181 | 0.0006237 |
| iiw6 | 6 | i1b | 1 | 90075 | 0.8542926 | 0.3133476 | 1.0000000 | 0.9333340 | 0.0000000 | 0.0232989 | 1.000000 | 9.767011e-01 | -1.8044064 | 1.4691605 | 0.0010441 |
| iiw7 | 7 | i1c | 1 | 30988 | 2.6599175 | 1.5977913 | 3.0455687 | 2.5897254 | 2.3898431 | 1.0000000 | 5.104169 | 4.104169e+00 | 0.1254937 | -1.7159442 | 0.0090766 |
| iiw8 | 8 | i1d | 1 | 30988 | 0.4845744 | 0.2113451 | 0.3917356 | 0.4541405 | 0.1221457 | 0.3093494 | 1.047275 | 7.379257e-01 | 0.8898299 | -0.4759514 | 0.0012006 |
| iiw9 | 9 | i2 | 1 | 90075 | 1.7202764 | 1.6227108 | 1.0000000 | 1.3104177 | 0.0000000 | 1.0000000 | 44.918862 | 4.391886e+01 | 2.6481313 | 13.3027889 | 0.0054068 |
| iiw10 | 10 | i3 | 1 | 90075 | 1.5860218 | 1.2752966 | 1.0000000 | 1.2569181 | 0.0000000 | 0.7554467 | 5.880294 | 5.124847e+00 | 1.9150319 | 2.0573721 | 0.0042492 |
| iiw11 | 11 | i3b | 1 | 90075 | 0.9053333 | 0.2172963 | 1.0000000 | 0.9668945 | 0.0000000 | 0.1634565 | 1.000000 | 8.365435e-01 | -2.2189636 | 3.5708295 | 0.0007240 |
| iiw12 | 12 | i4 | 1 | 90075 | 1.7202764 | 1.6227108 | 1.0000000 | 1.3104177 | 0.0000000 | 1.0000000 | 44.918862 | 4.391886e+01 | 2.6481313 | 13.3027889 | 0.0054068 |
| iiw13 | 13 | i4b | 1 | 90075 | 0.8535921 | 0.3146328 | 1.0000000 | 0.9326223 | 0.0000000 | 0.0086746 | 1.023063 | 1.014388e+00 | -1.7995183 | 1.4485057 | 0.0010483 |
| iiw14 | 14 | i4c | 1 | 30988 | 4.9086872 | 289.9031455 | 0.8565799 | 0.7783143 | 0.9844717 | 0.1925629 | 49523.911445 | 4.952372e+04 | 162.2555130 | 27512.6578355 | 1.6468580 |
| iiw15 | 15 | i4d | 1 | 30988 | 1.5153260 | 104.0729164 | 0.2103919 | 0.2517302 | 0.1181386 | 0.1307086 | 18136.282466 | 1.813615e+04 | 170.9329232 | 29750.5713178 | 0.5912089 |
| iiw16 | 16 | i5 | 1 | 90075 | 1.5860218 | 1.2752966 | 1.0000000 | 1.2569181 | 0.0000000 | 0.7554467 | 5.880294 | 5.124847e+00 | 1.9150319 | 2.0573721 | 0.0042492 |
| iiw17 | 17 | i5b | 1 | 90075 | 0.9053333 | 0.2172963 | 1.0000000 | 0.9668945 | 0.0000000 | 0.1634565 | 1.000000 | 8.365435e-01 | -2.2189636 | 3.5708295 | 0.0007240 |
Code
library(cowplot)Code
combined_plot <- plot_grid(ploti0, ploti0b, ploti0c, ploti0d, ploti1, ploti1b, ploti1c, ploti1d, ploti2, ploti3, ploti3b, ploti4, ploti4b, ploti4c, ploti4d, ploti5, ploti5b, ncol = 1, align = 'v', axis = 'tb')
print(combined_plot)
Code
ggsave(combined_plot, file="_proposal_grant/2023/_figs/comb_plot_24_04_18.jpg", width=8.5, height=15.5, dpi=500)Code
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
save.image(paste0(folder_path,"20240430_avance_iiw.RData"))We included several covariates in the model, such as PSU (policonsumo2), previous treatment completion status (if any, event.lag), previous biopsychosocial status (if any, lag(compromiso_biopsicosocial.y)), sex (sex), age of onset of substance use (edad_ini_cons), educational attainment (escolaridad_rec), primary substance at admission to treatment (sus_principal_mod), frequency of primary substance (freq_cons_sus_prin_ord), occupational status (condicion_ocupacional_corr24), number of children (binary, num_hijos_mod_joel_bin), housing tenure status (tenencia_de_la_vivienda_mod), administrative macro-zone of municipality of residence (macrozona), psychiatric co-morbidity (ICD-10, dg_cie_10_rec), SUD severity diagnosis (ICD-10, dg_trs_cons_sus_or), rurality status of municipality of residence (clas_r), poverty index of municipality of residence (porc_pobr), starting substance used (sus_ini_mod_mvv), corrected year of birth (ano_nac_corr), and cohabitation status (who the individual lives with, con_quien_vive_joel). These covariates are used to adjust for potential confounding factors that may affect the survival outcome.
We provisionally selected i4d (not only events; dichotomous transformation of previous biopsychosocial compromise, previous time in treatment (0-90 vs. >90), previous PSU and treatment counts). However, Lokku et al., does not recommend including treatment counts. So we ended selecting i1 : Not only events; dichotomous transformation of previous biopsychosocial compromise, previous time in treatment (0-90 vs. >90), previous PSU and lag first 1s; without treatment count.
Definitive model
To account for the dynamic relationship between time since follow-up, we included terms for time elapsed from the first treatment.
Code
invisible("2024-04-17: data_mine2_miss_proc2 not needed, data_mine_miss_proc2 instead (no longer needed to select events==1 [only with visits], because we already selected them)")
data_mine_miss_proc2 <-
dplyr::mutate(group_by(data_mine_miss_proc, hash_key), lag_tr_outcome= lag(tr_outcome, default=1), lag_comp_bpsc_y2_moderate= lag(comp_bpsc_y2_moderate, default=0), lag_comp_bpsc_y3_severe= lag(comp_bpsc_y3_severe, default=1), lag_fech_egres_num= lag(fech_egres_num), lag_less_90d_tr1= lag(less_90d_tr1, default=1), lag_policonsumo2= lag(policonsumo2, default=1), lag_dias_treat_imp_sin_na= lag(dias_treat_imp_sin_na)) %>% dplyr::ungroup()%>%
dplyr::mutate(log_lag_dias_treat_imp_sin_na= log(lag_dias_treat_imp_sin_na+0.001), log_dias_treat_imp_sin_na= log(dias_treat_imp_sin_na+0.001), esc_dum_rec_3prim = ifelse(as.character(escolaridad_rec)=="3-Completed primary school or less",1,0), esc_dum_rec_2high = ifelse(as.character(escolaridad_rec)== "2-Completed high school or less", 1, 0), susprindum_oh = ifelse(as.character(sus_principal_mod)== "Alcohol", 1, 0), susprindum_coc = ifelse(as.character(sus_principal_mod)== "Cocaine hydrochloride", 1, 0), susprindum_pbc = ifelse(as.character(sus_principal_mod)== "Cocaine paste", 1, 0), susprindum_mar = ifelse(as.character(sus_principal_mod)== "Marijuana", 1, 0), freq_cons_dum_5day= ifelse(as.character(freq_cons_sus_prin_ord)== "5. Daily", 1, 0), freq_cons_dum_44to6wk= ifelse(as.character(freq_cons_sus_prin_ord)== "4. 4 to 6 days a week", 1, 0), freq_cons_dum_32to3wk= ifelse(as.character(freq_cons_sus_prin_ord)== "3. 2 to 3 days a week", 1, 0), freq_cons_dum_21wkmore= ifelse(as.character(freq_cons_sus_prin_ord)== "2. 1 day a week or more", 1, 0), cond_oc_dum_3unemp= ifelse(as.character(condicion_ocupacional_corr24)== "Unemployed", 1, 0), cond_oc_dum_2inact= ifelse(as.character(condicion_ocupacional_corr24)== "Inactive", 1, 0), viv_dum_illegal= ifelse(as.character(tenencia_de_la_vivienda_mod)== "Illegal Settlement", 1, 0), viv_dum_own= ifelse(as.character(tenencia_de_la_vivienda_mod)== "Owner/Transferred dwellings/Pays Dividends", 1, 0), viv_dum_rent= ifelse(as.character(tenencia_de_la_vivienda_mod)== "Renting", 1, 0), viv_dum_temp= ifelse(as.character(tenencia_de_la_vivienda_mod)== "Stays temporarily with a relative", 1, 0), macro_dum_south= ifelse(as.character(macrozona)== "South", 1, 0), macro_dum_north= ifelse(as.character(macrozona)== "North", 1, 0), psycom_dum_with= ifelse(as.character(dg_cie_10_rec)== "With psychiatric comorbidity", 1, 0), psycom_dum_study= ifelse(as.character(dg_cie_10_rec)== "Diagnosis unknown (under study)", 1, 0), psycom_dum_study= ifelse(as.character(dg_cie_10_rec)== "Diagnosis unknown (under study)", 1, 0), rurality_rural= ifelse(as.character(clas_r)== "Rural", 1, 0), rurality_mix= ifelse(as.character(clas_r)== "Mixed", 1, 0), susinidum_oh = ifelse(as.character(sus_ini_mod_mvv)== "Alcohol", 1, 0), susinidum_coc = ifelse(as.character(sus_ini_mod_mvv)== "Cocaine hydrochloride", 1, 0), susinidum_pbc = ifelse(as.character(sus_ini_mod_mvv)== "Cocaine paste", 1, 0), susinidum_mar = ifelse(as.character(sus_ini_mod_mvv)== "Marijuana", 1, 0), cohab_dum_alone = ifelse(as.character(con_quien_vive_joel)== "Alone", 1, 0), cohab_dum_fam_or = ifelse(as.character(con_quien_vive_joel)== "Family of origin", 1, 0), cohab_dum_cpl_child = ifelse(as.character(con_quien_vive_joel)== "With couple/children", 1, 0), surv_time= ifelse(surv_time<0.001, 0.0001, surv_time)) %>% data.frame()
invisible("2024-05-02: problem with different values in people with plans (possibly imputed)")
data_mine_miss_restr_proc2 <-
dplyr::mutate(group_by(data_mine_miss_restr_proc, hash_key), lag_tr_outcome= lag(tr_outcome, default=1), lag_comp_bpsc_y2_moderate= lag(comp_bpsc_y2_moderate, default=0), lag_comp_bpsc_y3_severe= lag(comp_bpsc_y3_severe, default=1), lag_fech_egres_num= lag(fech_egres_num), lag_less_90d_tr1= lag(less_90d_tr1, default=1), lag_policonsumo2= lag(policonsumo2, default=1), lag_dias_treat_imp_sin_na= lag(dias_treat_imp_sin_na)) %>% dplyr::ungroup()%>%
dplyr::mutate(log_lag_dias_treat_imp_sin_na= log(lag_dias_treat_imp_sin_na+0.001), log_dias_treat_imp_sin_na= log(dias_treat_imp_sin_na+0.001), esc_dum_rec_3prim = ifelse(as.character(escolaridad_rec)=="3-Completed primary school or less",1,0), esc_dum_rec_2high = ifelse(as.character(escolaridad_rec)== "2-Completed high school or less", 1, 0), susprindum_oh = ifelse(as.character(sus_principal_mod)== "Alcohol", 1, 0), susprindum_coc = ifelse(as.character(sus_principal_mod)== "Cocaine hydrochloride", 1, 0), susprindum_pbc = ifelse(as.character(sus_principal_mod)== "Cocaine paste", 1, 0), susprindum_mar = ifelse(as.character(sus_principal_mod)== "Marijuana", 1, 0), freq_cons_dum_5day= ifelse(as.character(freq_cons_sus_prin_ord)== "5. Daily", 1, 0), freq_cons_dum_44to6wk= ifelse(as.character(freq_cons_sus_prin_ord)== "4. 4 to 6 days a week", 1, 0), freq_cons_dum_32to3wk= ifelse(as.character(freq_cons_sus_prin_ord)== "3. 2 to 3 days a week", 1, 0), freq_cons_dum_21wkmore= ifelse(as.character(freq_cons_sus_prin_ord)== "2. 1 day a week or more", 1, 0), cond_oc_dum_3unemp= ifelse(as.character(condicion_ocupacional_corr24)== "Unemployed", 1, 0), cond_oc_dum_2inact= ifelse(as.character(condicion_ocupacional_corr24)== "Inactive", 1, 0), viv_dum_illegal= ifelse(as.character(tenencia_de_la_vivienda_mod)== "Illegal Settlement", 1, 0), viv_dum_own= ifelse(as.character(tenencia_de_la_vivienda_mod)== "Owner/Transferred dwellings/Pays Dividends", 1, 0), viv_dum_rent= ifelse(as.character(tenencia_de_la_vivienda_mod)== "Renting", 1, 0), viv_dum_temp= ifelse(as.character(tenencia_de_la_vivienda_mod)== "Stays temporarily with a relative", 1, 0), macro_dum_south= ifelse(as.character(macrozona)== "South", 1, 0), macro_dum_north= ifelse(as.character(macrozona)== "North", 1, 0), psycom_dum_with= ifelse(as.character(dg_cie_10_rec)== "With psychiatric comorbidity", 1, 0), psycom_dum_study= ifelse(as.character(dg_cie_10_rec)== "Diagnosis unknown (under study)", 1, 0), psycom_dum_study= ifelse(as.character(dg_cie_10_rec)== "Diagnosis unknown (under study)", 1, 0), rurality_rural= ifelse(as.character(clas_r)== "Rural", 1, 0), rurality_mix= ifelse(as.character(clas_r)== "Mixed", 1, 0), susinidum_oh = ifelse(as.character(sus_ini_mod_mvv)== "Alcohol", 1, 0), susinidum_coc = ifelse(as.character(sus_ini_mod_mvv)== "Cocaine hydrochloride", 1, 0), susinidum_pbc = ifelse(as.character(sus_ini_mod_mvv)== "Cocaine paste", 1, 0), susinidum_mar = ifelse(as.character(sus_ini_mod_mvv)== "Marijuana", 1, 0), cohab_dum_alone = ifelse(as.character(con_quien_vive_joel)== "Alone", 1, 0), cohab_dum_fam_or = ifelse(as.character(con_quien_vive_joel)== "Family of origin", 1, 0), cohab_dum_cpl_child = ifelse(as.character(con_quien_vive_joel)== "With couple/children", 1, 0), surv_time= ifelse(surv_time<0.001, 0.0001, surv_time)) %>%
dplyr::group_by(id) %>%
dplyr::mutate(tipo_de_plan_2_mod=first(tipo_de_plan_2_mod)) %>%
dplyr::ungroup() %>%
data.frame()
dis_run_w_more_type_plan<-
data_mine_miss_restr_proc2 %>%
dplyr::group_by(id) %>%
dplyr::summarise(count_n_dis=n_distinct(tipo_de_plan_2_mod)) %>%
dplyr::filter(count_n_dis>1) %>% distinct(id)
# data_mine_miss_restr_proc2 %>%
# dplyr::group_by(id) %>%
# dplyr::mutate(tipo_de_plan_2_mod=first(tipo_de_plan_2_mod)) %>%
# dplyr::ungroup() %>%
# dplyr::filter(id %in% dis_run_w_more_type_plan$id)
invisible("export to python")
arrow::write_parquet(data_mine_miss_proc2 %>% dplyr::mutate(lag_time=ifelse(is.na(lag_time),0,lag_time)),
"_proposal_grant/2023/data_mine_miss_proc2_20240417.gz.parquet",
compression = "gzip", compression_level = 5)
arrow::write_parquet(data_mine_miss_restr_proc2 %>% dplyr::mutate(lag_time=ifelse(is.na(lag_time),0,lag_time)),
"_proposal_grant/2023/data_mine_miss_restr_proc2_20240427.gz.parquet",
compression = "gzip", compression_level = 5)
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Explore residuals")
coxph(Surv(lag_time,time,event)~ lag_tr_outcome+ lag_comp_bpsc_y3_severe+ lag_less_90d_tr1+ log_lag_dias_treat_imp_sin_na+ lag_policonsumo2+ edad_al_ing_1+ ano_nac_corr+ cluster(id)+ strata(tipo_de_plan_2), data = data_mine_miss_proc2)Call:
coxph(formula = Surv(lag_time, time, event) ~ lag_tr_outcome +
lag_comp_bpsc_y3_severe + lag_less_90d_tr1 + log_lag_dias_treat_imp_sin_na +
lag_policonsumo2 + edad_al_ing_1 + ano_nac_corr + strata(tipo_de_plan_2),
data = data_mine_miss_proc2, cluster = id)
coef exp(coef) se(coef) robust se z
lag_tr_outcome 0.139006 1.149130 0.019532 0.017338 8.017
lag_comp_bpsc_y3_severe 0.128602 1.137237 0.015722 0.014992 8.578
lag_less_90d_tr1 0.119770 1.127237 0.021018 0.023973 4.996
log_lag_dias_treat_imp_sin_na -0.027843 0.972541 0.007598 0.010508 -2.650
lag_policonsumo2 -0.011019 0.989041 0.020273 0.018189 -0.606
edad_al_ing_1 0.232879 1.262229 0.004098 0.003826 60.866
ano_nac_corr 0.236261 1.266505 0.004070 0.003788 62.373
p
lag_tr_outcome 1.08e-15
lag_comp_bpsc_y3_severe < 2e-16
lag_less_90d_tr1 5.85e-07
log_lag_dias_treat_imp_sin_na 0.00805
lag_policonsumo2 0.54462
edad_al_ing_1 < 2e-16
ano_nac_corr < 2e-16
Likelihood ratio test=3814 on 7 df, p=< 2.2e-16
n= 17671, number of events= 17671
(72404 observations deleted due to missingness)Code
# coef exp(coef) se(coef) robust se z p
# lag_tr_outcome 0.139016 1.149142 0.019531 0.017338 8.018 1.07e-15
# lag_comp_bpsc_y3_severe 0.129501 1.138261 0.015740 0.015006 8.630 < 2e-16
# lag_less_90d_tr1 0.119756 1.127222 0.021016 0.023976 4.995 5.89e-07
# log_lag_dias_treat_imp_sin_na -0.027876 0.972509 0.007595 0.010510 -2.652 0.008
# lag_policonsumo2 -0.010438 0.989616 0.020266 0.018185 -0.574 0.566
# edad_al_ing_1 0.232959 1.262330 0.004099 0.003827 60.870 < 2e-16
# ano_nac_corr 0.236313 1.266571 0.004070 0.003788 62.377 < 2e-16
coxph(Surv(lag_time,time,event)~ lag_tr_outcome+ lag_comp_bpsc_y3_severe+ lag_less_90d_tr1+ log_lag_dias_treat_imp_sin_na+ lag_policonsumo2+ edad_al_ing_1+ ano_nac_corr+ cluster(id)+ strata(tipo_de_plan_2), data = data_mine_miss_restr_proc2)Call:
coxph(formula = Surv(lag_time, time, event) ~ lag_tr_outcome +
lag_comp_bpsc_y3_severe + lag_less_90d_tr1 + log_lag_dias_treat_imp_sin_na +
lag_policonsumo2 + edad_al_ing_1 + ano_nac_corr + strata(tipo_de_plan_2),
data = data_mine_miss_restr_proc2, cluster = id)
coef exp(coef) se(coef) robust se z
lag_tr_outcome 0.139006 1.149130 0.019532 0.017338 8.017
lag_comp_bpsc_y3_severe 0.128602 1.137237 0.015722 0.014992 8.578
lag_less_90d_tr1 0.119770 1.127237 0.021018 0.023973 4.996
log_lag_dias_treat_imp_sin_na -0.027843 0.972541 0.007598 0.010508 -2.650
lag_policonsumo2 -0.011019 0.989041 0.020273 0.018189 -0.606
edad_al_ing_1 0.232879 1.262229 0.004098 0.003826 60.866
ano_nac_corr 0.236261 1.266505 0.004070 0.003788 62.373
p
lag_tr_outcome 1.08e-15
lag_comp_bpsc_y3_severe < 2e-16
lag_less_90d_tr1 5.85e-07
log_lag_dias_treat_imp_sin_na 0.00805
lag_policonsumo2 0.54462
edad_al_ing_1 < 2e-16
ano_nac_corr < 2e-16
Likelihood ratio test=3814 on 7 df, p=< 2.2e-16
n= 17671, number of events= 17671
(13317 observations deleted due to missingness)Code
# coef exp(coef) se(coef) robust se z p
# lag_tr_outcome 0.139 1.149 0.020 0.017 8.0 1e-15
# lag_comp_bpsc_y3_severe 0.128 1.137 0.016 0.015 8.6 <2e-16
# lag_less_90d_tr1 0.120 1.128 0.021 0.024 5.0 5e-07
# log_lag_dias_treat_imp_sin_na -0.028 0.972 0.008 0.011 -2.7 0.008
# lag_policonsumo2 -0.010 0.990 0.020 0.018 -0.6 0.574
# edad_al_ing_1 0.233 1.262 0.004 0.004 60.9 <2e-16
# ano_nac_corr 0.236 1.267 0.004 0.004 62.4 <2e-16
cox.zph(
coxph(Surv(lag_time,time,event)~ lag_tr_outcome+ lag_comp_bpsc_y3_severe+ lag_less_90d_tr1+ log_lag_dias_treat_imp_sin_na+ lag_policonsumo2+ edad_al_ing_1+ ano_nac_corr+ cluster(id)+ strata(tipo_de_plan_2), data= data_mine_miss_proc2)
) chisq df p
lag_tr_outcome 37.88 1 7.5e-10
lag_comp_bpsc_y3_severe 61.07 1 5.5e-15
lag_less_90d_tr1 147.34 1 < 2e-16
log_lag_dias_treat_imp_sin_na 178.31 1 < 2e-16
lag_policonsumo2 13.27 1 0.00027
edad_al_ing_1 35.36 1 2.7e-09
ano_nac_corr 8.64 1 0.00328
GLOBAL 521.18 7 < 2e-16Code
# chisq df p
# lag_tr_outcome 37.98 1 7.1e-10
# lag_comp_bpsc_y3_severe 61.87 1 3.7e-15
# lag_less_90d_tr1 147.28 1 < 2e-16
# log_lag_dias_treat_imp_sin_na 178.13 1 < 2e-16
# lag_policonsumo2 13.18 1 0.00028
# edad_al_ing_1 35.57 1 2.5e-09
# ano_nac_corr 8.73 1 0.00312
# GLOBAL 521.92 7 < 2e-16
cox.zph(
coxph(Surv(lag_time,time,event)~ lag_tr_outcome+ lag_comp_bpsc_y3_severe+ lag_less_90d_tr1+ log_lag_dias_treat_imp_sin_na+ lag_policonsumo2+ edad_al_ing_1+ ano_nac_corr+ cluster(id)+ strata(tipo_de_plan_2), data= data_mine_miss_restr_proc2)
) chisq df p
lag_tr_outcome 37.88 1 7.5e-10
lag_comp_bpsc_y3_severe 61.07 1 5.5e-15
lag_less_90d_tr1 147.34 1 < 2e-16
log_lag_dias_treat_imp_sin_na 178.31 1 < 2e-16
lag_policonsumo2 13.27 1 0.00027
edad_al_ing_1 35.36 1 2.7e-09
ano_nac_corr 8.64 1 0.00328
GLOBAL 521.18 7 < 2e-16Code
# chisq df p
# lag_tr_outcome 37.95 1 7.3e-10
# lag_comp_bpsc_y3_severe 57.93 1 2.7e-14
# lag_less_90d_tr1 147.25 1 < 2e-16
# log_lag_dias_treat_imp_sin_na 178.15 1 < 2e-16
# lag_policonsumo2 13.21 1 0.00028
# edad_al_ing_1 35.50 1 2.6e-09
# ano_nac_corr 8.71 1 0.00317
# GLOBAL 519.01 7 < 2e-16
invisible("Tiene mejor ajuste que las alternativas sólo con días de tto. continuos (log) o el dicotomizado solo: 186751.2 vs. 186756.6 vs. 186779.1; 2024-04-21: 261769.5 el actual sin moderado 262293.3, vs. mucho menos cuando intento agregar términos. Pero lo haré dirigido por la teoría")
cox.zph(
coxph(Surv(lag_time,time,event)~ lag_tr_outcome+ lag_comp_bpsc_y3_severe+ lag_less_90d_tr1+ log_lag_dias_treat_imp_sin_na+ lag_policonsumo2+ edad_al_ing_1+ ano_nac_corr+ cluster(id)+ strata(tipo_de_plan_2_mod), data= data_mine_miss_restr_proc2)
) chisq df p
lag_tr_outcome 54.28 1 1.7e-13
lag_comp_bpsc_y3_severe 20.29 1 6.7e-06
lag_less_90d_tr1 130.84 1 < 2e-16
log_lag_dias_treat_imp_sin_na 142.09 1 < 2e-16
lag_policonsumo2 9.78 1 0.0018
edad_al_ing_1 33.55 1 7.0e-09
ano_nac_corr 7.91 1 0.0049
GLOBAL 460.60 7 < 2e-16Code
# chisq df p
# lag_tr_outcome 54.66 1 1.4e-13
# lag_comp_bpsc_y3_severe 19.67 1 9.2e-06
# lag_less_90d_tr1 130.48 1 < 2e-16
# log_lag_dias_treat_imp_sin_na 141.52 1 < 2e-16
# lag_policonsumo2 9.72 1 0.0018
# edad_al_ing_1 33.47 1 7.2e-09
# ano_nac_corr 7.88 1 0.0050
# GLOBAL 459.34 7 < 2e-16
final <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ log_dias_treat_imp_sin_na.lag+ policonsumo2.lag+ edad_al_ing_1+ ano_nac_corr+
cluster(hash_key),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc2,
invariant= c("id", "edad_al_ing_1", "ano_nac_corr"),
lagvars= c("time", "tr_outcome", "comp_bpsc_y3_severe", "less_90d_tr1", "log_dias_treat_imp_sin_na","policonsumo2"),
maxfu= maxfu_df,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(2.95082,1,1,1,4.499811,1), #90/30.5 4.499811 es 90 días
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
final_restr <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ log_dias_treat_imp_sin_na.lag+ policonsumo2.lag+ edad_al_ing_1+ ano_nac_corr+
cluster(hash_key),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_restr_proc2,
invariant= c("id", "edad_al_ing_1", "ano_nac_corr"),
lagvars= c("time", "tr_outcome", "comp_bpsc_y3_severe", "less_90d_tr1", "log_dias_treat_imp_sin_na","policonsumo2"),
maxfu= maxfu_restr_df,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(2.95082,1,1,1,4.499811,1), #90/30.5 4.499811 es 90 días
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
final_b <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ log_dias_treat_imp_sin_na.lag+ policonsumo2.lag+ edad_al_ing_1+ ano_nac_corr+
cluster(hash_key),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_proc2,
invariant= c("id", "edad_al_ing_1", "ano_nac_corr"),
lagvars= c("time", "tr_outcome", "comp_bpsc_y3_severe", "less_90d_tr1", "log_dias_treat_imp_sin_na","policonsumo2"),
maxfu= maxfu_df,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(2.95082,0,0,0,4.499811/2,0), #90/30.5 4.499811 es 90 días
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
final_restr_b <- iiw.weights(
Surv(time.lag,time,event)~ tr_outcome.lag+ comp_bpsc_y3_severe.lag+ less_90d_tr1.lag+ log_dias_treat_imp_sin_na.lag+ policonsumo2.lag+ edad_al_ing_1+ ano_nac_corr+
cluster(hash_key),
id= "id",
time= "time",
event= "event",
data= data_mine_miss_restr_proc2,
invariant= c("id", "edad_al_ing_1", "ano_nac_corr"),
lagvars= c("time", "tr_outcome", "comp_bpsc_y3_severe", "less_90d_tr1", "log_dias_treat_imp_sin_na","policonsumo2"),
maxfu= maxfu_restr_df,#Base_fiscalia_v15f_grant_23_24_long2$cens_time,
lagfirst= c(2.95082,0,0,0,4.499811/2,0), #90/30.5 4.499811 es 90 días
first= T #If TRUE, the first observation for each individual is assigned an intensity of 1. This is appropriate if the first visit is a baseline visit at which recruitment to the study occurred
)
#
broom::tidy(final$m, conf.int=T, exponentiate =T)# A tibble: 7 x 8
term estim~1 std.e~2 robus~3 stati~4 p.value conf.~5 conf.~6
<chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 1.33 0.0194 0.0183 15.6 4.62e- 55 1.28 1.38
2 comp_bpsc_y3_severe~ 1.72 0.0129 0.0135 40.3 0 1.68 1.77
3 less_90d_tr1.lag 3.30 0.0179 0.0205 58.3 0 3.17 3.44
4 log_dias_treat_imp_~ 1.39 0.0106 0.0139 23.9 2.11e-126 1.36 1.43
5 policonsumo2.lag 1.10 0.0191 0.0196 5.07 4.03e- 7 1.06 1.15
6 edad_al_ing_1 1.08 0.00153 0.00141 57.9 0 1.08 1.09
7 ano_nac_corr 1.09 0.00153 0.00140 64.4 0 1.09 1.10
# ... with abbreviated variable names 1: estimate, 2: std.error, 3: robust.se,
# 4: statistic, 5: conf.low, 6: conf.highCode
# term estimate std.error robust.se statistic p.value conf.low conf.high
# <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
# 1 tr_outcome.lag 1.17 0.0189 0.0172 9.01 2.07e-19 1.13 1.21
# 2 comp_bpsc_y3_severe.lag 1.80 0.0126 0.0130 45.3 0 1.75 1.85
# 3 less_90d_tr1.lag 2.27 0.0130 0.0142 57.6 0 2.21 2.34
# 4 policonsumo2.lag 1.18 0.0189 0.0196 8.30 1.06e-16 1.13 1.22
# 5 edad_al_ing_1 1.09 0.00153 0.00139 60.3 0 1.08 1.09
# 6 ano_nac_corr 1.10 0.00153 0.00138 66.2 0 1.09 1.10
broom::tidy(final_restr$m, conf.int=T, exponentiate =T)# A tibble: 7 x 8
term estim~1 std.e~2 robus~3 stati~4 p.value conf.~5 conf.~6
<chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 1.27 0.0196 0.0175 13.7 5.54e- 43 1.23 1.32
2 comp_bpsc_y3_severe~ 1.58 0.0135 0.0130 35.0 4.07e-268 1.54 1.62
3 less_90d_tr1.lag 2.47 0.0198 0.0221 40.9 0 2.37 2.58
4 log_dias_treat_imp_~ 1.26 0.0109 0.0132 17.6 1.17e- 69 1.23 1.29
5 policonsumo2.lag 1.12 0.0195 0.0179 6.17 6.73e- 10 1.08 1.16
6 edad_al_ing_1 1.15 0.00296 0.00266 53.2 0 1.15 1.16
7 ano_nac_corr 1.15 0.00293 0.00263 54.4 0 1.15 1.16
# ... with abbreviated variable names 1: estimate, 2: std.error, 3: robust.se,
# 4: statistic, 5: conf.low, 6: conf.highCode
# term estimate std.error robust.se statistic p.value conf.low conf.high
# <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
# tr_outcome.lag 1.27 0.0196 0.0175 13.7 9.41e- 43 1.23 1.32
# comp_bpsc_y3_severe.lag 1.58 0.0135 0.0130 35.1 5.67e-270 1.54 1.62
# less_90d_tr1.lag 2.47 0.0198 0.0221 41.0 0 2.37 2.58
# log_dias_treat_imp_sin_na.lag 1.26 0.0109 0.0132 17.6 1.97e- 69 1.23 1.29
# policonsumo2.lag 1.12 0.0195 0.0179 6.31 2.70e- 10 1.08 1.16
# edad_al_ing_1 1.15 0.00296 0.00267 53.2 0 1.15 1.16
# ano_nac_corr 1.15 0.00293 0.00263 54.4 0 1.15 1.16
broom::tidy(final_b$m, conf.int=T, exponentiate =T)# A tibble: 7 x 8
term estim~1 std.e~2 robus~3 stati~4 p.value conf.~5 conf.~6
<chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 0.661 0.0144 0.0197 -21.0 5.52e- 98 0.636 0.687
2 comp_bpsc_y3_severe~ 0.913 0.0144 0.0193 -4.72 2.40e- 6 0.879 0.948
3 less_90d_tr1.lag 0.871 0.0180 0.0248 -5.58 2.38e- 8 0.829 0.914
4 log_dias_treat_imp_~ 0.829 0.00303 0.0121 -15.5 4.29e- 54 0.810 0.849
5 policonsumo2.lag 0.531 0.0144 0.0239 -26.5 4.80e-155 0.507 0.556
6 edad_al_ing_1 1.07 0.00153 0.00140 50.1 0 1.07 1.08
7 ano_nac_corr 1.09 0.00152 0.00140 58.9 0 1.08 1.09
# ... with abbreviated variable names 1: estimate, 2: std.error, 3: robust.se,
# 4: statistic, 5: conf.low, 6: conf.highCode
# term estimate std.error robust.se statistic p.value conf.low conf.high
# <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
# 1 tr_outcome.lag 0.551 0.0133 0.0152 -39.3 0 0.534 0.567
# 2 comp_bpsc_y3_severe.lag 0.866 0.0142 0.0162 -8.88 6.86e-19 0.839 0.894
# 3 less_90d_tr1.lag 0.904 0.0184 0.0263 -3.83 1.31e- 4 0.859 0.952
# 4 log_dias_treat_imp_sin_na.lag 0.873 0.00421 0.0137 -9.86 6.23e-23 0.850 0.897
# 5 policonsumo2.lag 0.450 0.0135 0.0167 -47.9 0 0.435 0.465
# 6 edad_al_ing_1 1.07 0.00153 0.00138 51.0 0 1.07 1.08
# 7 ano_nac_corr 1.09 0.00152 0.00138 60.0 0 1.08 1.09
broom::tidy(final_restr_b$m, conf.int=T, exponentiate =T)# A tibble: 7 x 8
term estim~1 std.e~2 robus~3 stati~4 p.value conf.~5 conf.~6
<chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 tr_outcome.lag 0.679 0.0150 0.0185 -21.0 1.78e- 97 0.654 0.704
2 comp_bpsc_y3_severe~ 0.937 0.0146 0.0178 -3.67 2.39e- 4 0.904 0.970
3 less_90d_tr1.lag 0.857 0.0181 0.0246 -6.29 3.10e- 10 0.816 0.899
4 log_dias_treat_imp_~ 0.835 0.00326 0.0130 -13.9 6.21e- 44 0.814 0.857
5 policonsumo2.lag 0.576 0.0152 0.0211 -26.1 2.67e-150 0.552 0.600
6 edad_al_ing_1 1.14 0.00293 0.00274 47.3 0 1.13 1.14
7 ano_nac_corr 1.15 0.00293 0.00268 52.1 0 1.14 1.16
# ... with abbreviated variable names 1: estimate, 2: std.error, 3: robust.se,
# 4: statistic, 5: conf.low, 6: conf.highCode
# term estimate std.error robust.se statistic p.value conf.low conf.high
# <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
# tr_outcome.lag 0.603 0.0144 0.0155 -32.6 1.17e-232 0.585 0.621
# comp_bpsc_y3_severe.lag 0.913 0.0144 0.0149 -6.14 8.03e- 10 0.886 0.940
# less_90d_tr1.lag 0.898 0.0187 0.0249 -4.34 1.45e- 5 0.855 0.943
# log_dias_treat_imp_sin_na.lag 0.887 0.00464 0.0134 -8.92 4.53e- 19 0.864 0.911
# policonsumo2.lag 0.514 0.0146 0.0163 -40.9 0 0.498 0.530
# edad_al_ing_1 1.13 0.00291 0.00250 48.1 0 1.12 1.13
# ano_nac_corr 1.14 0.00291 0.00249 52.9 0 1.14 1.15
final_coefs<-
cbind.data.frame(term = c("tr_outcome.lag", "comp_bpsc_y3_severe.lag", "less_90d_tr1.lag", "policonsumo2.lag", "edad_al_ing_1", "ano_nac_corr"), estimate = c(1.1680161176676, 1.79963517024553, 2.2710236051536, 1.17641492065206, 1.08756180792251, 1.09565232375642), std.error = c(0.0188883080181889, 0.0125981011876289, 0.0129873997440068, 0.0189035437658918, 0.00153425515290826, 0.00153498904364182), robust.se = c(0.0172377882900819, 0.0129754816518389, 0.0142405606352216, 0.0195792679568791, 0.00139105857932987, 0.00138071917329718), statistic = c(9.0096641790833, 45.2841734070737, 57.5981998354651, 8.29814532853555, 60.3413245741051, 66.1611116269822), p.value = c(2.06688423909932e-19, 0, 0, 1.05749268471728e-16, 0, 0), conf.low = c(1.12921334955458, 1.75444485412689, 2.20851351433301, 1.13212562626954, 1.0846006909948, 1.09269132170511), conf.high = c(1.20815225189237, 1.84598948115495, 2.33530299076412, 1.22243683335131, 1.09053100912818, 1.09862134960454))
summary(final$iiw) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.07435 1.00000 1.00000 0.90727 1.00000 68.81131 Code
#
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.1 1.0 1.0 0.9 1.0 67.9
summary(final_restr$iiw) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.08549 0.44872 0.84983 0.74315 1.00000 22.25216 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.09 0.45 0.85 0.74 1.00 22.04
invisible("Generar los pesos a asignar")
table(final$iiw, data_mine_miss_proc$treatment) %>% data.frame() %>% dplyr::mutate(Var1=as.numeric(as.character(Var1)), Freq=as.numeric(Freq)) %>% dplyr::arrange(-Freq) %>% dplyr::select(-Var2) %>% head(20) %>% dput()structure(list(Var1 = c(1, 0.2934922955593, 0.0743489610317217,
0.0757106912282225, 0.0762091661974078, 0.0783327140965704, 0.0803105805175816,
0.0803344064102367, 0.0804539791818909, 0.0820904544526977, 0.0841827879338751,
0.0844743814169596, 0.0849411523170872, 0.0851903256041754, 0.0855473256157542,
0.0865205249520326, 0.0867570352679417, 0.088066938352575, 0.0880950865189643,
0.0881021667861931), Freq = c(72404, 2, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1)), row.names = c(NA, 20L), class = "data.frame")Code
df_iiw_final<-
cbind.data.frame(Var1 = c(1, 0.688125981472609, 0.160330998755955, 0.162315144547036, 0.196774517463519, 0.199606787793155, 0.208958220652199, 0.225731677712565, 0.23160770335121, 0.234224555582056, 0.248919669741449, 0.256310292435046, 0.275606737385485, 0.286468644357305, 0.291310136239186, 0.313585733564029, 0.314758256721433, 0.325438651308499, 0.329390866243409,
0.334785138867613), Freq = c(72404, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2))
invisible("Vemos un ejemplo que los pesos asignados son ABSOLUTAMENTE DISTINTOS DEPENDIENDO DEL MODELO")
cbind.data.frame(mod1= final$iiw, mod2= final_b$iiw) %>% dplyr::filter(mod1!=mod2) %>% head(20) mod1 mod2
1 0.1996962 3.830127
2 0.6129391 4.376810
3 0.4980109 5.125454
4 0.2382301 4.547194
5 0.5663714 3.425793
6 0.2268115 4.153721
7 0.1720342 5.678893
8 0.4969182 5.333688
9 0.1529589 3.750321
10 0.6696665 3.895218
11 0.5430455 2.042192
12 0.3793116 3.575538
13 0.4025839 3.457417
14 0.3729438 4.203298
15 0.4674447 2.872873
16 0.2808815 2.449836
17 0.5428789 8.560529
18 0.4052071 8.143959
19 0.7632280 7.063683
20 0.5840069 1.996160Code
diff_in_iiw_final<-
cbind.data.frame(mod1 = c(0.199498133334888, 0.615044327911636, 0.498702140851273, 0.238887400402573, 0.564883027201795, 0.226856508027775, 0.172052897285973, 0.495624040479963, 0.152900128488306, 0.669574361459713, 0.543818575958686, 0.379205064180558, 0.402370920088421, 0.373500499352896, 0.467980694989343, 0.281860985585125, 0.544134768464911, 0.404800245361857, 0.763907890441244, 0.585275800391092), mod2 = c(3.92318595976502, 3.44378094343031, 4.97127471370872, 4.5599185879251, 3.04873855278048,
3.31979843463701, 6.0865617052278, 5.81549682361537, 4.09528203750446, 3.99883088697191, 1.7257302695626, 3.11919908361139, 3.04433093376514, 4.05407333418453, 2.28706611143663, 2.08872102380155, 8.02426306030948, 8.36617898514813, 6.98296103354168, 1.33559630246317))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
require(rms)
cox_iiw_proposed<-
rms::cph(Surv(lag_time,time,event)~
cluster(id)+
lag_tr_outcome+
lag_less_90d_tr1+
log_lag_dias_treat_imp_sin_na+
lag_comp_bpsc_y3_severe+
lag_policonsumo2+
edad_al_ing_1+
ano_nac_corr+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
psycom_dum_with+
psycom_dum_study+
freq_cons_dum_5day+
cond_oc_dum_2inact+
cond_oc_dum_3unemp+
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar +
strat(tipo_de_plan_2_mod),
data = data_mine_miss_proc2,
x=TRUE,
y=TRUE,
surv= TRUE)
#Must use x=TRUE,y=TRUE to get survival curves with time-dep. covariables
cox_iiw_proposedFrequencies of Missing Values Due to Each Variable
Surv(lag_time, time, event) cluster(id)
72404 0
lag_tr_outcome lag_less_90d_tr1
0 0
log_lag_dias_treat_imp_sin_na lag_comp_bpsc_y3_severe
72404 0
lag_policonsumo2 edad_al_ing_1
0 0
ano_nac_corr susinidum_oh
0 0
susinidum_coc susinidum_pbc
0 0
susinidum_mar psycom_dum_with
0 0
psycom_dum_study freq_cons_dum_5day
0 0
cond_oc_dum_2inact cond_oc_dum_3unemp
0 0
susprindum_oh susprindum_coc
0 0
susprindum_pbc susprindum_mar
0 0
tipo_de_plan_2_mod
0
Cox Proportional Hazards Model
rms::cph(formula = Surv(lag_time, time, event) ~ cluster(id) +
lag_tr_outcome + lag_less_90d_tr1 + log_lag_dias_treat_imp_sin_na +
lag_comp_bpsc_y3_severe + lag_policonsumo2 + edad_al_ing_1 +
ano_nac_corr + susinidum_oh + susinidum_coc + susinidum_pbc +
susinidum_mar + psycom_dum_with + psycom_dum_study + freq_cons_dum_5day +
cond_oc_dum_2inact + cond_oc_dum_3unemp + susprindum_oh +
susprindum_coc + susprindum_pbc + susprindum_mar + strat(tipo_de_plan_2_mod),
data = data_mine_miss_proc2, x = TRUE, y = TRUE, surv = TRUE)
Status
Stratum No Event Event
tipo_de_plan_2_mod=basic ambulatory 0 5622
tipo_de_plan_2_mod=GP intensive ambulatory 0 6501
tipo_de_plan_2_mod=GP residential 0 3034
tipo_de_plan_2_mod=WO intensive ambulatory 0 1015
tipo_de_plan_2_mod=WO residential 0 1499
Model Tests Discrimination
Indexes
Obs 17671 LR chi2 3826.50 R2 0.195
Events 17671 d.f. 20 R2(20,17671)0.194
Center 474.9181 Pr(> chi2) 0.0000 Dxy 0.255
Score chi2 3855.18
Pr(> chi2) 0.0000
Coef S.E. Wald Z Pr(>|Z|)
lag_tr_outcome 0.1545 0.0176 8.76 <0.0001
lag_less_90d_tr1 0.1033 0.0242 4.27 <0.0001
log_lag_dias_treat_imp_sin_na -0.0222 0.0100 -2.22 0.0263
lag_comp_bpsc_y3_severe 0.0584 0.0165 3.54 0.0004
lag_policonsumo2 -0.0181 0.0186 -0.97 0.3314
edad_al_ing_1 0.2345 0.0039 60.56 <0.0001
ano_nac_corr 0.2359 0.0038 61.97 <0.0001
susinidum_oh 0.0441 0.0465 0.95 0.3436
susinidum_coc 0.1096 0.0573 1.91 0.0558
susinidum_pbc 0.1586 0.0527 3.01 0.0026
susinidum_mar 0.1589 0.0474 3.35 0.0008
psycom_dum_with 0.0174 0.0161 1.08 0.2783
psycom_dum_study 0.0304 0.0224 1.36 0.1741
freq_cons_dum_5day 0.0135 0.0155 0.87 0.3834
cond_oc_dum_2inact 0.0581 0.0214 2.72 0.0066
cond_oc_dum_3unemp 0.0621 0.0171 3.63 0.0003
susprindum_oh -0.1040 0.0626 -1.66 0.0968
susprindum_coc -0.1105 0.0636 -1.74 0.0822
susprindum_pbc -0.0906 0.0622 -1.46 0.1452
susprindum_mar -0.1151 0.0688 -1.67 0.0945
Code
# Cox Proportional Hazards Model
#
# rms::cph(formula = Surv(lag_time, time, event) ~ lag_tr_outcome +
# lag_less_90d_tr1 + log_lag_dias_treat_imp_sin_na + lag_comp_bpsc_y3_severe +
# lag_policonsumo2 + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
# susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar + cluster(id) + strat(tipo_de_plan_2_mod),
# data = data_mine_miss_proc2, x = TRUE, y = TRUE, surv = TRUE)
#
#
# Status
# Stratum No Event Event
# tipo_de_plan_2_mod=basic ambulatory 0 5622
# tipo_de_plan_2_mod=GP intensive ambulatory 0 6501
# tipo_de_plan_2_mod=GP residential 0 3034
# tipo_de_plan_2_mod=WO intensive ambulatory 0 1015
# tipo_de_plan_2_mod=WO residential 0 1499
#
# Model Tests Discrimination
# Indexes
# Obs 17671 LR chi2 3826.50 R2 0.195
# Events 17671 d.f. 20 R2(20,17671)0.194
# Center 474.9181 Pr(> chi2) 0.0000 Dxy 0.255
# Score chi2 3855.18
# Pr(> chi2) 0.0000
#
# Coef S.E. Wald Z Pr(>|Z|)
# lag_tr_outcome 0.1545 0.0176 8.76 <0.0001
# lag_less_90d_tr1 0.1033 0.0242 4.27 <0.0001
# log_lag_dias_treat_imp_sin_na -0.0222 0.0100 -2.22 0.0263
# lag_comp_bpsc_y3_severe 0.0584 0.0165 3.54 0.0004
# lag_policonsumo2 -0.0181 0.0186 -0.97 0.3314
# edad_al_ing_1 0.2345 0.0039 60.56 <0.0001
# ano_nac_corr 0.2359 0.0038 61.97 <0.0001
# susinidum_oh 0.0441 0.0465 0.95 0.3436
# susinidum_coc 0.1096 0.0573 1.91 0.0558
# susinidum_pbc 0.1586 0.0527 3.01 0.0026
# susinidum_mar 0.1589 0.0474 3.35 0.0008
# psycom_dum_with 0.0174 0.0161 1.08 0.2783
# psycom_dum_study 0.0304 0.0224 1.36 0.1741
# freq_cons_dum_5day 0.0135 0.0155 0.87 0.3834
# cond_oc_dum_2inact 0.0581 0.0214 2.72 0.0066
# cond_oc_dum_3unemp 0.0621 0.0171 3.63 0.0003
# susprindum_oh -0.1040 0.0626 -1.66 0.0968
# susprindum_coc -0.1105 0.0636 -1.74 0.0822
# susprindum_pbc -0.0906 0.0622 -1.46 0.1452
# susprindum_mar -0.1151 0.0688 -1.67 0.0945
round(cox.zph(cox_iiw_proposed)$table, 4) chisq df p
lag_tr_outcome 53.1085 1 0.0000
lag_less_90d_tr1 130.0606 1 0.0000
log_lag_dias_treat_imp_sin_na 142.4076 1 0.0000
lag_comp_bpsc_y3_severe 20.3681 1 0.0000
lag_policonsumo2 8.3301 1 0.0039
edad_al_ing_1 33.2395 1 0.0000
ano_nac_corr 7.7556 1 0.0054
susinidum_oh 40.0796 1 0.0000
susinidum_coc 0.9605 1 0.3271
susinidum_pbc 6.6977 1 0.0097
susinidum_mar 24.9223 1 0.0000
psycom_dum_with 16.2090 1 0.0001
psycom_dum_study 37.2938 1 0.0000
freq_cons_dum_5day 2.1105 1 0.1463
cond_oc_dum_2inact 3.1074 1 0.0779
cond_oc_dum_3unemp 14.3746 1 0.0001
susprindum_oh 24.9308 1 0.0000
susprindum_coc 0.7316 1 0.3924
susprindum_pbc 19.7982 1 0.0000
susprindum_mar 2.6082 1 0.1063
GLOBAL 495.5153 20 0.0000Code
# lag_tr_outcome 53.1085 1 0.0000 ###
# lag_less_90d_tr1 130.0606 1 0.0000 ###
# log_lag_dias_treat_imp_sin_na 142.4076 1 0.0000 ###
# lag_comp_bpsc_y3_severe 20.3681 1 0.0000 ###
# lag_policonsumo2 8.3301 1 0.0039 #
# edad_al_ing_1 33.2395 1 0.0000 ###
# ano_nac_corr 7.7556 1 0.0054 #
# susinidum_oh 40.0796 1 0.0000 ###
# susinidum_coc 0.9605 1 0.3271
# susinidum_pbc 6.6977 1 0.0097
# susinidum_mar 24.9223 1 0.0000 ###
# psycom_dum_with 16.2090 1 0.0001 ###
# psycom_dum_study 37.2938 1 0.0000 ###
# freq_cons_dum_5day 2.1105 1 0.1463
# cond_oc_dum_2inact 3.1074 1 0.0779
# cond_oc_dum_3unemp 14.3746 1 0.0001 ###
# susprindum_oh 24.9308 1 0.0000 ###
# susprindum_coc 0.7316 1 0.3924
# susprindum_pbc 19.7982 1 0.0000 ###
# susprindum_mar 2.6082 1 0.1063
# GLOBAL 495.5153 20 0.0000
cox_iiw_restr_proposed<-
rms::cph(Surv(lag_time,time,event)~
cluster(id)+
lag_tr_outcome+
lag_less_90d_tr1+
log_lag_dias_treat_imp_sin_na+
lag_comp_bpsc_y3_severe+
lag_policonsumo2+
edad_al_ing_1+
ano_nac_corr+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
psycom_dum_with+
psycom_dum_study+
freq_cons_dum_5day+
cond_oc_dum_2inact+
cond_oc_dum_3unemp+
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar +
strat(tipo_de_plan_2_mod),
data = data_mine_miss_restr_proc2,
x=TRUE,
y=TRUE,
surv= TRUE)
#Must use x=TRUE,y=TRUE to get survival curves with time-dep. covariables
cox_iiw_restr_proposedFrequencies of Missing Values Due to Each Variable
Surv(lag_time, time, event) cluster(id)
13317 0
lag_tr_outcome lag_less_90d_tr1
0 0
log_lag_dias_treat_imp_sin_na lag_comp_bpsc_y3_severe
13317 0
lag_policonsumo2 edad_al_ing_1
0 0
ano_nac_corr susinidum_oh
0 0
susinidum_coc susinidum_pbc
0 0
susinidum_mar psycom_dum_with
0 0
psycom_dum_study freq_cons_dum_5day
0 0
cond_oc_dum_2inact cond_oc_dum_3unemp
0 0
susprindum_oh susprindum_coc
0 0
susprindum_pbc susprindum_mar
0 0
tipo_de_plan_2_mod
0
Cox Proportional Hazards Model
rms::cph(formula = Surv(lag_time, time, event) ~ cluster(id) +
lag_tr_outcome + lag_less_90d_tr1 + log_lag_dias_treat_imp_sin_na +
lag_comp_bpsc_y3_severe + lag_policonsumo2 + edad_al_ing_1 +
ano_nac_corr + susinidum_oh + susinidum_coc + susinidum_pbc +
susinidum_mar + psycom_dum_with + psycom_dum_study + freq_cons_dum_5day +
cond_oc_dum_2inact + cond_oc_dum_3unemp + susprindum_oh +
susprindum_coc + susprindum_pbc + susprindum_mar + strat(tipo_de_plan_2_mod),
data = data_mine_miss_restr_proc2, x = TRUE, y = TRUE, surv = TRUE)
Status
Stratum No Event Event
tipo_de_plan_2_mod=basic ambulatory 0 5633
tipo_de_plan_2_mod=GP intensive ambulatory 0 6499
tipo_de_plan_2_mod=GP residential 0 3026
tipo_de_plan_2_mod=WO intensive ambulatory 0 1015
tipo_de_plan_2_mod=WO residential 0 1498
Model Tests Discrimination
Indexes
Obs 17671 LR chi2 3826.06 R2 0.195
Events 17671 d.f. 20 R2(20,17671)0.194
Center 474.8266 Pr(> chi2) 0.0000 Dxy 0.255
Score chi2 3854.74
Pr(> chi2) 0.0000
Coef S.E. Wald Z Pr(>|Z|)
lag_tr_outcome 0.1540 0.0176 8.73 <0.0001
lag_less_90d_tr1 0.1036 0.0242 4.28 <0.0001
log_lag_dias_treat_imp_sin_na -0.0223 0.0100 -2.23 0.0257
lag_comp_bpsc_y3_severe 0.0590 0.0165 3.58 0.0003
lag_policonsumo2 -0.0181 0.0186 -0.97 0.3312
edad_al_ing_1 0.2345 0.0039 60.55 <0.0001
ano_nac_corr 0.2359 0.0038 61.96 <0.0001
susinidum_oh 0.0441 0.0465 0.95 0.3434
susinidum_coc 0.1097 0.0573 1.91 0.0555
susinidum_pbc 0.1589 0.0527 3.01 0.0026
susinidum_mar 0.1589 0.0474 3.35 0.0008
psycom_dum_with 0.0174 0.0161 1.09 0.2778
psycom_dum_study 0.0303 0.0224 1.35 0.1762
freq_cons_dum_5day 0.0139 0.0155 0.89 0.3716
cond_oc_dum_2inact 0.0581 0.0214 2.72 0.0065
cond_oc_dum_3unemp 0.0625 0.0171 3.66 0.0003
susprindum_oh -0.1038 0.0626 -1.66 0.0973
susprindum_coc -0.1104 0.0636 -1.74 0.0826
susprindum_pbc -0.0902 0.0622 -1.45 0.1467
susprindum_mar -0.1150 0.0688 -1.67 0.0947
Code
# Cox Proportional Hazards Model
#
# rms::cph(formula = Surv(lag_time, time, event) ~ lag_tr_outcome +
# lag_less_90d_tr1 + log_lag_dias_treat_imp_sin_na + lag_comp_bpsc_y3_severe +
# lag_policonsumo2 + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
# susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar + cluster(id) + strat(tipo_de_plan_2_mod),
# data = data_mine_miss_restr_proc2, x = TRUE, y = TRUE, surv = TRUE)
#
#
# Status
# Stratum No Event Event
# tipo_de_plan_2_mod=basic ambulatory 0 5622
# tipo_de_plan_2_mod=GP intensive ambulatory 0 6501
# tipo_de_plan_2_mod=GP residential 0 3034
# tipo_de_plan_2_mod=WO intensive ambulatory 0 1015
# tipo_de_plan_2_mod=WO residential 0 1499
#
# Model Tests Discrimination
# Indexes
# Obs 17671 LR chi2 3826.50 R2 0.195
# Events 17671 d.f. 20 R2(20,17671)0.194
# Center 474.9181 Pr(> chi2) 0.0000 Dxy 0.255
# Score chi2 3855.18
# Pr(> chi2) 0.0000
#
# Coef S.E. Wald Z Pr(>|Z|)
# lag_tr_outcome 0.1545 0.0176 8.76 <0.0001
# lag_less_90d_tr1 0.1033 0.0242 4.27 <0.0001
# log_lag_dias_treat_imp_sin_na -0.0222 0.0100 -2.22 0.0263
# lag_comp_bpsc_y3_severe 0.0584 0.0165 3.54 0.0004
# lag_policonsumo2 -0.0181 0.0186 -0.97 0.3314
# edad_al_ing_1 0.2345 0.0039 60.56 <0.0001
# ano_nac_corr 0.2359 0.0038 61.97 <0.0001
# susinidum_oh 0.0441 0.0465 0.95 0.3436
# susinidum_coc 0.1096 0.0573 1.91 0.0558
# susinidum_pbc 0.1586 0.0527 3.01 0.0026
# susinidum_mar 0.1589 0.0474 3.35 0.0008
# psycom_dum_with 0.0174 0.0161 1.08 0.2783
# psycom_dum_study 0.0304 0.0224 1.36 0.1741
# freq_cons_dum_5day 0.0135 0.0155 0.87 0.3834
# cond_oc_dum_2inact 0.0581 0.0214 2.72 0.0066
# cond_oc_dum_3unemp 0.0621 0.0171 3.63 0.0003
# susprindum_oh -0.1040 0.0626 -1.66 0.0968
# susprindum_coc -0.1105 0.0636 -1.74 0.0822
# susprindum_pbc -0.0906 0.0622 -1.46 0.1452
# susprindum_mar -0.1151 0.0688 -1.67 0.0945
cox.zph(cox_iiw_restr_proposed) chisq df p
lag_tr_outcome 52.745 1 3.8e-13
lag_less_90d_tr1 130.440 1 < 2e-16
log_lag_dias_treat_imp_sin_na 143.001 1 < 2e-16
lag_comp_bpsc_y3_severe 20.975 1 4.7e-06
lag_policonsumo2 8.383 1 0.00379
edad_al_ing_1 33.324 1 7.8e-09
ano_nac_corr 7.786 1 0.00526
susinidum_oh 40.366 1 2.1e-10
susinidum_coc 0.960 1 0.32728
susinidum_pbc 6.850 1 0.00887
susinidum_mar 24.998 1 5.7e-07
psycom_dum_with 16.153 1 5.8e-05
psycom_dum_study 37.275 1 1.0e-09
freq_cons_dum_5day 2.262 1 0.13261
cond_oc_dum_2inact 3.092 1 0.07866
cond_oc_dum_3unemp 14.720 1 0.00012
susprindum_oh 25.145 1 5.3e-07
susprindum_coc 0.758 1 0.38382
susprindum_pbc 20.116 1 7.3e-06
susprindum_mar 2.568 1 0.10906
GLOBAL 497.106 20 < 2e-16Code
# chisq df p
# lag_tr_outcome 53.109 1 3.2e-13
# lag_less_90d_tr1 130.061 1 < 2e-16
# log_lag_dias_treat_imp_sin_na 142.408 1 < 2e-16
# lag_comp_bpsc_y3_severe 20.368 1 6.4e-06
# lag_policonsumo2 8.330 1 0.00390
# edad_al_ing_1 33.239 1 8.1e-09
# ano_nac_corr 7.756 1 0.00535
# susinidum_oh 40.080 1 2.4e-10
# susinidum_coc 0.961 1 0.32706
# susinidum_pbc 6.698 1 0.00965
# susinidum_mar 24.922 1 6.0e-07
# psycom_dum_with 16.209 1 5.7e-05
# psycom_dum_study 37.294 1 1.0e-09
# freq_cons_dum_5day 2.110 1 0.14630
# cond_oc_dum_2inact 3.107 1 0.07794
# cond_oc_dum_3unemp 14.375 1 0.00015
# susprindum_oh 24.931 1 5.9e-07
# susprindum_coc 0.732 1 0.39237
# susprindum_pbc 19.798 1 8.6e-06
# susprindum_mar 2.608 1 0.10631
# GLOBAL 495.515 20 < 2e-16
invisible("552.545, schoenfeld residuals; now, 496")
res <- resid(cox_iiw_proposed, "scaledsch")
time <- as.numeric(dimnames(res)[[1]])
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
jpeg(filename = paste0(folder_path,"zph/schoefeld_res_all_variables_240421.jpg"), width = 8, height = 6, units = "in", res = 600, quality = 600)#Users/CISS Fondecyt/
par(mfrow = c(3, 3))
z2 <- loess(res[,"lag_tr_outcome"] ~ time, span=0.50) # residuals
plot(time, fitted(z2))
lines(supsmu(time, res[,"lag_tr_outcome"]),lty=2)
title("Sch. res., lag_tr_outcome")
z3 <- loess(res[,"lag_less_90d_tr1"] ~ time, span=0.50) # residual
plot(time, fitted(z3))
lines(supsmu(time, res[,"lag_less_90d_tr1"]),lty=2)
title("Sch. res., lag_less_90d_tr1")
z4 <- loess(res[,"log_lag_dias_treat_imp_sin_na"] ~ time, span=0.50) # residuals
plot(time, fitted(z4))
lines(supsmu(time, res[,"log_lag_dias_treat_imp_sin_na"]),lty=2)
title("Sch. res., log_lag_dias_treat_imp_sin_na")
z5 <- loess(res[,"lag_comp_bpsc_y3_severe"] ~ time, span=0.50) # residuals
plot(time, fitted(z5))
lines(supsmu(time, res[,"lag_comp_bpsc_y3_severe"]),lty=2)
title("Sch. res., lag_comp_bpsc_y3_severe")
z6 <- loess(res[,"lag_policonsumo2"] ~ time, span=0.50) # residuals
plot(time, fitted(z6))
lines(supsmu(time, res[,"lag_policonsumo2"]),lty=2)
title("Sch. res., lag_policonsumo2")
z7 <- loess(res[,"edad_al_ing_1"] ~ time, span=0.50) # residuals
plot(time, fitted(z7))
lines(supsmu(time, res[,"edad_al_ing_1"]),lty=2)
title("Sch. res., edad_al_ing_1")
z8 <- loess(res[,"susinidum_oh"] ~ time, span=0.50) # residuals
plot(time, fitted(z8))
lines(supsmu(time, res[,"susinidum_oh"]),lty=2)
title("Sch. res., susinidum_oh")
z9 <- loess(res[,"psycom_dum_study"] ~ time, span=0.50) # residuals
plot(time, fitted(z9))
lines(supsmu(time, res[,"psycom_dum_study"]),lty=2)
title("Sch. res., psycom_dum_study")
dev.off()png
2 Code
#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_}
#_#_#_#_#_#_#_#_#_#_#_}
invisible("Explore residuals in restricted subsample")
#_#_#_#_#_#_#_#_#_#_#_}#_#_#_#_#_#_#_#_#_#_#_}
res_restr <- resid(cox_iiw_restr_proposed, "scaledsch")
time_restr <- as.numeric(dimnames(res_restr)[[1]])
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
jpeg(filename = paste0(folder_path,"zph/schoefeld_res_all_variables_restr_240428.jpg"), width = 8, height = 6, units = "in", res = 600, quality = 600)#Users/CISS Fondecyt/
par(mfrow = c(3, 3))
z2_restr <- loess(res_restr[,"lag_tr_outcome"] ~ time, span=0.50) # residuals
plot(time_restr, fitted(z2_restr))
lines(supsmu(time_restr, res_restr[,"lag_tr_outcome"]),lty=2)
title("Sch. res., lag_tr_outcome")
z3_restr <- loess(res_restr[,"lag_less_90d_tr1"] ~ time, span=0.50) # residual
plot(time_restr, fitted(z3_restr))
lines(supsmu(time_restr, res_restr[,"lag_less_90d_tr1"]),lty=2)
title("Sch. res., lag_less_90d_tr1")
z4_restr <- loess(res_restr[,"log_lag_dias_treat_imp_sin_na"] ~ time, span=0.50) # residuals
plot(time, fitted(z4_restr))
lines(supsmu(time_restr, res_restr[,"log_lag_dias_treat_imp_sin_na"]),lty=2)
title("Sch. res., log_lag_dias_treat_imp_sin_na")
z5_restr <- loess(res_restr[,"lag_comp_bpsc_y3_severe"] ~ time, span=0.50) # residuals
plot(time_restr, fitted(z5_restr))
lines(supsmu(time_restr, res_restr[,"lag_comp_bpsc_y3_severe"]),lty=2)
title("Sch. res., lag_comp_bpsc_y3_severe")
z6_restr <- loess(res_restr[,"lag_policonsumo2"] ~ time, span=0.50) # residuals
plot(time_restr, fitted(z6_restr))
lines(supsmu(time_restr, res_restr[,"lag_policonsumo2"]),lty=2)
title("Sch. res., lag_policonsumo2")
z7_restr <- loess(res_restr[,"edad_al_ing_1"] ~ time, span=0.50) # residuals
plot(time_restr, fitted(z7_restr))
lines(supsmu(time_restr, res_restr[,"edad_al_ing_1"]),lty=2)
title("Sch. res., edad_al_ing_1")
z8_restr <- loess(res_restr[,"susinidum_oh"] ~ time, span=0.50) # residuals
plot(time_restr, fitted(z8_restr))
lines(supsmu(time_restr, res_restr[,"susinidum_oh"]),lty=2)
title("Sch. res., susinidum_oh")
z9_restr <- loess(res_restr[,"psycom_dum_study"] ~ time, span=0.50) # residuals
plot(time_restr, fitted(z9_restr))
lines(supsmu(time_restr, res_restr[,"psycom_dum_study"]),lty=2)
title("Sch. res., psycom_dum_study")
dev.off()png
2 Code
#save.image("C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/20240422_avance_iiw.RData")
#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_}
plot(cox.zph(cox_iiw_proposed)[which(attr(cox_iiw_proposed$coefficients, "names") == "log_lag_dias_treat_imp_sin_na")], lwd=2)
abline(0,0, col="red", lty=3, lwd=2)
abline(h=cox_iiw_proposed$coefficients["log_lag_dias_treat_imp_sin_na"], col=3, lwd=2, lty= 3)
legend("bottomleft", legend= c("Reference line for the null effect", "Average hazard over time", "Time-varying hazard"), lty=c(3,2,1), col= c("red",3,1), lwd=2)
Code
#_#_#_#_#_#_#_#_#_#_#_}
km_days_tr <- survfit(Surv(lag_time,time,event)~ log_lag_dias_treat_imp_sin_na+
cluster(id), data = data_mine_miss_proc2)
plot(km_days_tr, fun = "cloglog", xlab = "Time (in days) using log",
ylab = "log-log survival", main = "log-log curves by lag_less_90d_tr1")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1)
Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Restricted")
plot(cox.zph(cox_iiw_restr_proposed)[which(attr(cox_iiw_restr_proposed$coefficients, "names") == "log_lag_dias_treat_imp_sin_na")], lwd=2)
abline(0,0, col="red", lty=3, lwd=2)
abline(h=cox_iiw_restr_proposed$coefficients["log_lag_dias_treat_imp_sin_na"], col=3, lwd=2, lty= 3)
legend("bottomleft", legend= c("Reference line for the null effect", "Average hazard over time", "Time-varying hazard"), lty=c(3,2,1), col= c("red",3,1), lwd=2)
Code
#_#_#_#_#_#_#_#_#_#_#_}
km_days_tr_restr <- survfit(Surv(lag_time,time,event)~ log_lag_dias_treat_imp_sin_na+
cluster(id), data = data_mine_miss_restr_proc2)
plot(km_days_tr_restr, fun = "cloglog", xlab = "Time (in days) using log",
ylab = "log-log survival", main = "log-log curves by log_lag_dias_treat_imp_sin_na (restr)")
Code
#legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1)Transforming variables
Code
#load("C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/20240422_avance_iiw.RData")
#load("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/20240422_avance_iiw.RData")
require(splines)Code
invisible("Primero necesito buscar log_lag_dias_treat_imp_sin_na , luego lag_comp_bpsc_y3_severe, edad_al_ing_1 y porc_pobr")
#specs_dias_treat <- paste0("log_lag_dias_treat_imp_sin_na *", c("time", "poly(time, 3)", "poly(time, 4)", "poly(time, 5)", "poly(time, 6)", "bs(time, 3)", "bs(time, 4)", "bs(time, 5)", "bs(time, 6)", "ns(time, 3)", "ns(time, 4)", "ns(time, 5)", "ns(time, 6)", "rcs(time, 3)", "rcs(time, 4)", "rcs(time, 5)", "rcs(time, 6)"))
specs_dias_treat <- paste0("", c("lag_dias_treat_imp_sin_na", "poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 3)", "poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 4)", "poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 5)", "poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 6)", "bs(log_lag_dias_treat_imp_sin_na, 3)", "bs(log_lag_dias_treat_imp_sin_na, 4)", "bs(log_lag_dias_treat_imp_sin_na, 5)", "bs(log_lag_dias_treat_imp_sin_na, 6)", "ns(log_lag_dias_treat_imp_sin_na, 3)", "ns(log_lag_dias_treat_imp_sin_na, 4)", "ns(log_lag_dias_treat_imp_sin_na, 5)", "ns(log_lag_dias_treat_imp_sin_na, 6)", "rcs(log_lag_dias_treat_imp_sin_na, 3)", "rcs(log_lag_dias_treat_imp_sin_na, 4)", "rcs(log_lag_dias_treat_imp_sin_na, 5)", "rcs(log_lag_dias_treat_imp_sin_na, 6)"))
aic_values0 <- vector(length=length(specs_dias_treat))
coxzph_values0 <-vector(length=length(specs_dias_treat))
for (i in seq_along(specs_dias_treat)) {
l <- specs_dias_treat[i]
# TryCatch block to handle errors during model fitting
#result <- tryCatch({
model <- cph(as.formula(paste("Surv(lag_time,time,event==1)~
cluster(id) +
lag_tr_outcome +
lag_less_90d_tr1 +",
l,
"lag_comp_bpsc_y3_severe +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar +
strat(tipo_de_plan_2)", sep=" +")),
data=data_mine_miss_proc2, x=TRUE, y=TRUE, surv=TRUE)
aic_val <- extractAIC(model)[[2]]
cox_zph_val <- cox.zph(model)$table[nrow(cox.zph(model)$table), 1]
# return(c(aic_val, cox_zph_val))
# }, error = function(e) {
# return(c(NA, NA))
# })
aic_values0[i] <- aic_val
coxzph_values0[i] <- cox_zph_val
print(cat(paste("Spec: ", l, "\n")))
print(cat(paste("AIC: ", aic_val[1], "\n")))
print(cat(paste("Cox zph test statistic: ", cox_zph_val[1], "\n")))
}Spec: lag_dias_treat_imp_sin_na
NULL
AIC: 258772.431420878
NULL
Cox zph test statistic: 700.698782419001
NULL
Spec: poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 3)
NULL
AIC: 258752.758407148
NULL
Cox zph test statistic: 681.320358260574
NULL
Spec: poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 4)
NULL
AIC: 258752.352037703
NULL
Cox zph test statistic: 685.49092454129
NULL
Spec: poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 5)
NULL
AIC: 258754.237212064
NULL
Cox zph test statistic: 688.359919424507
NULL
Spec: poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 6)
NULL
AIC: 258756.065355281
NULL
Cox zph test statistic: 699.028740090731
NULL
Spec: bs(log_lag_dias_treat_imp_sin_na, 3)
NULL
AIC: 258752.758407148
NULL
Cox zph test statistic: 681.32035826058
NULL
Spec: bs(log_lag_dias_treat_imp_sin_na, 4)
NULL
AIC: 258752.591529064
NULL
Cox zph test statistic: 692.29531570604
NULL
Spec: bs(log_lag_dias_treat_imp_sin_na, 5)
NULL
AIC: 258754.375773626
NULL
Cox zph test statistic: 692.851794860437
NULL
Spec: bs(log_lag_dias_treat_imp_sin_na, 6)
NULL
AIC: 258755.643123221
NULL
Cox zph test statistic: 692.08808495201
NULL
Spec: ns(log_lag_dias_treat_imp_sin_na, 3)
NULL
AIC: 258751.283635165
NULL
Cox zph test statistic: 686.265288516322
NULL
Spec: ns(log_lag_dias_treat_imp_sin_na, 4)
NULL
AIC: 258753.142150926
NULL
Cox zph test statistic: 689.263524577162
NULL
Spec: ns(log_lag_dias_treat_imp_sin_na, 5)
NULL
AIC: 258754.209648194
NULL
Cox zph test statistic: 688.709704889273
NULL
Spec: ns(log_lag_dias_treat_imp_sin_na, 6)
NULL
AIC: 258755.749792632
NULL
Cox zph test statistic: 688.462498438466
NULL
Spec: rcs(log_lag_dias_treat_imp_sin_na, 3)
NULL
AIC: 258770.719152328
NULL
Cox zph test statistic: 684.053264675061
NULL
Spec: rcs(log_lag_dias_treat_imp_sin_na, 4)
NULL
AIC: 258769.093325505
NULL
Cox zph test statistic: 690.115532063767
NULL
Spec: rcs(log_lag_dias_treat_imp_sin_na, 5)
NULL
AIC: 258770.713482787
NULL
Cox zph test statistic: 693.692481658824
NULL
Spec: rcs(log_lag_dias_treat_imp_sin_na, 6)
NULL
AIC: 258769.353464904
NULL
Cox zph test statistic: 697.798435975237
NULLCode
message(paste0("There was no model that improved cox zph of the restricted sample (vs. ", round(cox.zph(cox_iiw_proposed)$table[nrow(cox.zph(cox_iiw_proposed)$table), 1],1),")"))Code
#There was no model that improved cox zph of the restricted sample (vs. 495.5)
aic_values0_restr <- vector(length=length(specs_dias_treat))
coxzph_values0_restr <-vector(length=length(specs_dias_treat))
for (i in seq_along(specs_dias_treat)) {
l <- specs_dias_treat[i]
# TryCatch block to handle errors during model fitting
#result <- tryCatch({
model <- cph(as.formula(paste("Surv(lag_time,time,event==1)~
cluster(id) +
lag_tr_outcome +
lag_less_90d_tr1 +",
"I(round(",l,",2))+",
"lag_comp_bpsc_y3_severe +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar +
strat(tipo_de_plan_2_mod)", sep="")),
data=data_mine_miss_restr_proc2, x=TRUE, y=TRUE, surv=TRUE)
aic_val_restr <- extractAIC(model)[[2]]
cox_zph_val_restr <- cox.zph(model)$table[nrow(cox.zph(model)$table), 1]
# return(c(aic_val, cox_zph_val))
# }, error = function(e) {
# return(c(NA, NA))
# })
aic_values0_restr[i] <- aic_val_restr
coxzph_values0_restr[i] <- cox_zph_val_restr
print(cat(paste("Spec: ", l, "\n")))
print(cat(paste("AIC: ", aic_val_restr[1], "\n")))
print(cat(paste("Cox zph test statistic: ", cox_zph_val_restr[1], "\n")))
}Spec: lag_dias_treat_imp_sin_na
NULL
AIC: 246485.806233321
NULL
Cox zph test statistic: 639.107234814594
NULL
Spec: poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 3)
NULL
AIC: 246485.395254038
NULL
Cox zph test statistic: 563.585278389179
NULL
Spec: poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 4)
NULL
AIC: 246487.221628619
NULL
Cox zph test statistic: 567.377829244832
NULL
Spec: poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 5)
NULL
AIC: 246487.489808795
NULL
Cox zph test statistic: 589.036438983096
NULL
Spec: poly(ifelse(is.na(log_lag_dias_treat_imp_sin_na),.001,log_lag_dias_treat_imp_sin_na), 6)
NULL
AIC: 246489.392166329
NULL
Cox zph test statistic: 594.368973890416
NULL
Spec: bs(log_lag_dias_treat_imp_sin_na, 3)
NULL
AIC: 246471.938352736
NULL
Cox zph test statistic: 619.854794379174
NULL
Spec: bs(log_lag_dias_treat_imp_sin_na, 4)
NULL
AIC: 246471.919669293
NULL
Cox zph test statistic: 630.581400491839
NULL
Spec: bs(log_lag_dias_treat_imp_sin_na, 5)
NULL
AIC: 246473.415412343
NULL
Cox zph test statistic: 632.455278206685
NULL
Spec: bs(log_lag_dias_treat_imp_sin_na, 6)
NULL
AIC: 246473.86359796
NULL
Cox zph test statistic: 631.440180459019
NULL
Spec: ns(log_lag_dias_treat_imp_sin_na, 3)
NULL
AIC: 246471.745534296
NULL
Cox zph test statistic: 625.638986339057
NULL
Spec: ns(log_lag_dias_treat_imp_sin_na, 4)
NULL
AIC: 246471.324101404
NULL
Cox zph test statistic: 626.922786275545
NULL
Spec: ns(log_lag_dias_treat_imp_sin_na, 5)
NULL
AIC: 246473.612068326
NULL
Cox zph test statistic: 626.854990209742
NULL
Spec: ns(log_lag_dias_treat_imp_sin_na, 6)
NULL
AIC: 246475.935992451
NULL
Cox zph test statistic: 627.023802589597
NULL
Spec: rcs(log_lag_dias_treat_imp_sin_na, 3)
NULL
AIC: 246481.982378402
NULL
Cox zph test statistic: 625.011096881098
NULL
Spec: rcs(log_lag_dias_treat_imp_sin_na, 4)
NULL
AIC: 246479.729583509
NULL
Cox zph test statistic: 629.005973316837
NULL
Spec: rcs(log_lag_dias_treat_imp_sin_na, 5)
NULL
AIC: 246480.719358582
NULL
Cox zph test statistic: 633.068535007454
NULL
Spec: rcs(log_lag_dias_treat_imp_sin_na, 6)
NULL
AIC: 246480.612072818
NULL
Cox zph test statistic: 631.710467477473
NULLCode
message(paste0("There was no model that improved cox zph of the restricted sample (vs. ", round(cox.zph(cox_iiw_restr_proposed)$table[nrow(cox.zph(cox_iiw_restr_proposed)$table), 1],1),")"))Code
#There was no model that improved cox zph of the restricted sample (vs. 495.5)
model_after_mod_log_lag_dias_treat<-
cph(Surv(lag_time,time,event)~
cluster(id)+
lag_tr_outcome +
log_lag_dias_treat_imp_sin_na_rec.1+
log_lag_dias_treat_imp_sin_na_rec.2+
log_lag_dias_treat_imp_sin_na_rec.3+
lag_less_90d_tr1+
lag_comp_bpsc_y3_severe +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar +
strat(tipo_de_plan_2),
data= data_mine_miss_proc2 %>% dplyr::mutate(log_lag_dias_treat_imp_sin_na_rec= bs(log_lag_dias_treat_imp_sin_na, 3)) %>% data.table::as.data.table() %>% data.frame(), x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_after_mod_log_lag_dias_treat) chisq df p
lag_tr_outcome 42.103 1 8.7e-11
log_lag_dias_treat_imp_sin_na_rec.1 239.412 1 < 2e-16
log_lag_dias_treat_imp_sin_na_rec.2 275.169 1 < 2e-16
log_lag_dias_treat_imp_sin_na_rec.3 310.127 1 < 2e-16
lag_less_90d_tr1 165.596 1 < 2e-16
lag_comp_bpsc_y3_severe 59.927 1 9.8e-15
lag_policonsumo2 11.242 1 0.00080
edad_al_ing_1 38.030 1 7.0e-10
ano_nac_corr 9.980 1 0.00158
susinidum_oh 48.294 1 3.7e-12
susinidum_coc 0.687 1 0.40706
susinidum_pbc 11.416 1 0.00073
susinidum_mar 27.299 1 1.7e-07
psycom_dum_with 11.212 1 0.00081
psycom_dum_study 42.523 1 7.0e-11
freq_cons_dum_5day 14.890 1 0.00011
cond_oc_dum_2inact 0.241 1 0.62313
cond_oc_dum_3unemp 28.081 1 1.2e-07
susprindum_oh 35.197 1 3.0e-09
susprindum_coc 1.806 1 0.17901
susprindum_pbc 33.775 1 6.2e-09
susprindum_mar 1.108 1 0.29243
GLOBAL 681.320 22 < 2e-16Code
invisible("The unadjusted was 547; now its 673 21 < 2e-16; we kept the simplest model")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Now we see less than 90 days in treatment interactions, given that the interaction with time of log tr. days did not improved the model")
par(mfrow=c(3,1))
par(mar = c(5, 4, 4, 5))
res240422_1 <- resid(cox_iiw_proposed, "scaledsch")
time240422_1 <- as.numeric(dimnames(res240422_1)[[1]])
z_240422_1_1 <- loess(res240422_1[,"lag_less_90d_tr1"] ~ time240422_1, span=0.50) # residuals
# x11()
plot(time240422_1, fitted(z_240422_1_1))
lines(supsmu(time240422_1, res240422_1[,"psycom_dum_study"]),lty=2)
title("Sch. res., lag_less_90d_tr1")
plot(cox.zph(cox_iiw_proposed)[which(attr(cox_iiw_proposed$coefficients, "names") == "lag_less_90d_tr1")], lwd=2)
abline(0,0, col="red", lty=3, lwd=2)
abline(h=cox_iiw_proposed$coefficients["lag_less_90d_tr1"], col=3, lwd=2, lty= 3)
legend("bottomleft", legend= c("Reference line for the null effect", "Average hazard over time", "Time-varying hazard"), lty=c(3,2,1), col= c("red",3,1), lwd=2, bty = "n", bg = "transparent")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
km_home <- survfit(Surv(lag_time,time,event)~ lag_less_90d_tr1+
cluster(id), data = data_mine_miss_proc2)
#autoplot(km_home) # just to see the km curves
plot(km_home, fun = "cloglog", xlab = "Time (in days) using log",col=c("black", "red"),
ylab = "log-log survival", main = "log-log curves by lag_less_90d_tr1")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1, bty = "n", bg = "transparent")
Code
recorded_plot <- recordPlot()
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
png(paste0(folder_path,"zph/proportionallity_lag_less_90d_tr1.png"), height=13, width=10, res=500, units="in")
recorded_plot
dev.off()png
2 Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("restricted")
par(mfrow=c(2,1))
par(mar = c(5, 4, 4, 5))
plot(cox.zph(cox_iiw_restr_proposed)[which(attr(cox_iiw_restr_proposed$coefficients, "names") == "lag_less_90d_tr1")], lwd=2)
abline(0,0, col="red", lty=3, lwd=2)
abline(h=cox_iiw_restr_proposed$coefficients["lag_less_90d_tr1"], col=3, lwd=2, lty= 3)
legend("bottomleft", legend= c("Reference line for the null effect", "Average hazard over time", "Time-varying hazard"), lty=c(3,2,1), col= c("red",3,1), lwd=2, bty = "n", bg = "transparent")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
km_less90d_restr <- survfit(Surv(lag_time,time,event)~ lag_less_90d_tr1+
cluster(id), data = data_mine_miss_restr_proc2)
#autoplot(km_home) # just to see the km curves
plot(km_less90d_restr, fun = "cloglog", xlab = "Time (in days) using log",col=c("black", "red"),
ylab = "log-log survival", main = "log-log curves by lag_less_90d_tr1 (restr)")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1, bty = "n", bg = "transparent")
Code
recorded_plot <- recordPlot()
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
png(paste0(folder_path,"zph/proportionallity_lag_less_90d_tr1_restr.png"), height=13, width=10, res=500, units="in")
recorded_plot
dev.off()png
2 Code
create_time_varying_effect <- function(binary_covariate, t, decrease_end, increase_start, max_time) {
# Initialize the effect to be equal to the covariate
adjusted_effect <- binary_covariate
# Decrease phase: a sharp decrease from the start until decrease_end
# The covariate's effect is reduced to zero by decrease_end
adjusted_effect <- ifelse(t <= decrease_end, binary_covariate * (decrease_end - t) / decrease_end, adjusted_effect)
# Flat phase: between decrease_end and increase_start, the effect is zero
adjusted_effect <- ifelse(t > decrease_end & t <= increase_start, 0, adjusted_effect)
# Increase phase: a gradual increase from increase_start to max_time
# The covariate's effect is increased back to its original value by max_time
time_range <- max_time - increase_start
adjusted_effect <- ifelse(t > increase_start, binary_covariate * (t - increase_start) / time_range, adjusted_effect)
return(adjusted_effect)
}
data_mine_miss_proc2$lag_less_90d_tr1_rec<-
create_time_varying_effect(data_mine_miss_proc2$lag_less_90d_tr1,
data_mine_miss_proc2$time,
20,
80,
max(data_mine_miss_proc2$time))
data_mine_miss_restr_proc2$lag_less_90d_tr1_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$lag_less_90d_tr1,
data_mine_miss_restr_proc2$time,
20,
80,
max(data_mine_miss_restr_proc2$time))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Remove residuals")
mean_resid <- mean(res240422_1[,"lag_less_90d_tr1"], na.rm = TRUE)
sd_resid <- sd(res240422_1[,"lag_less_90d_tr1"], na.rm = TRUE)
# Identify outliers as those residuals greater than 3 standard deviations from the mean
outliers <- which(abs(res240422_1[,"lag_less_90d_tr1"] - mean_resid) > (3 * sd_resid))
data_clean <- data_mine_miss_proc2[-outliers, ]
invisible("even it worse results. Not important //DISCARDED")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Used a clog log plot, and told chat gpt to suggest a function")
#Here's a possible function that you can tailor to fit the divergence you observe in your log-log plot:
time_varying_coeff <- function(t, change_points, coefficients) {
# t: vector of times
# change_points: vector of times at which the hazard ratio changes
# coefficients: vector of coefficients corresponding to intervals defined by change_points
# Ensure change_points are sorted
change_points <- sort(c(0, change_points, Inf))
# Create a vector to hold the coefficient for each time t
t_coeff <- rep(0, length(t))
# Loop over the intervals defined by change_points
for (i in seq_along(coefficients)) {
interval_start <- change_points[i]
interval_end <- change_points[i + 1]
# Assign the coefficient to times within the current interval
t_coeff[t > interval_start & t <= interval_end] <- coefficients[i]
}
return(t_coeff)
}
# Define the change points and coefficients based on your data and log-log plot
change_points <- c(10, 20, 30) # Example change points (months)
coefficients <- c(0.5, 1.5, 0.5, 1) # Example coefficients
# Add the time-varying coefficient to your model
data_mine_miss_proc2$lag_less_90d_tr1_rec2 <- with(data_mine_miss_proc2,
lag_less_90d_tr1*time_varying_coeff(lag_time, change_points, coefficients))
data_mine_miss_proc2$less_90d_tr1_rec2 <- with(data_mine_miss_proc2,
less_90d_tr1*time_varying_coeff(time, change_points, coefficients))
invisible("2024-04-30: restricted")
data_mine_miss_restr_proc2$lag_less_90d_tr1_rec2 <- with(data_mine_miss_restr_proc2,
lag_less_90d_tr1*time_varying_coeff(lag_time, change_points, coefficients))
data_mine_miss_restr_proc2$less_90d_tr1_rec2 <- with(data_mine_miss_restr_proc2,
less_90d_tr1*time_varying_coeff(time, change_points, coefficients))
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
model_after_mod_not_log_lag_dias_treat_lag_less_90d<-
cph(Surv(lag_time,time,event)~
cluster(id)+
lag_tr_outcome +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec2+
lag_comp_bpsc_y3_severe +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2),
#data= data_clean,
data= data_mine_miss_proc2 %>% dplyr::mutate(log_lag_dias_treat_imp_sin_na_rec= bs(log_lag_dias_treat_imp_sin_na, 3)) %>% data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_after_mod_not_log_lag_dias_treat_lag_less_90d) chisq df p
lag_tr_outcome 37.918 1 7.4e-10
log_lag_dias_treat_imp_sin_na 179.021 1 < 2e-16
lag_less_90d_tr1_rec2 89.397 1 < 2e-16
lag_comp_bpsc_y3_severe 59.737 1 1.1e-14
lag_policonsumo2 10.980 1 0.00092
edad_al_ing_1 35.338 1 2.8e-09
ano_nac_corr 8.684 1 0.00321
susinidum_oh 45.623 1 1.4e-11
susinidum_coc 0.773 1 0.37918
susinidum_pbc 10.736 1 0.00105
susinidum_mar 25.726 1 3.9e-07
psycom_dum_with 10.019 1 0.00155
psycom_dum_study 42.363 1 7.6e-11
freq_cons_dum_5day 15.415 1 8.6e-05
cond_oc_dum_2inact 0.113 1 0.73706
cond_oc_dum_3unemp 27.499 1 1.6e-07
susprindum_oh 32.538 1 1.2e-08
susprindum_coc 1.707 1 0.19134
susprindum_pbc 31.291 1 2.2e-08
susprindum_mar 1.017 1 0.31319
GLOBAL 543.671 20 < 2e-16Code
# chisq df p
# log_lag_dias_treat_imp_sin_na 163.562 1 < 2e-16
# lag_less_90d_tr1_rec2 85.349 1 < 2e-16
# lag_comp_bpsc_y3_severe 55.509 1 9.3e-14
# lag_policonsumo2 11.327 1 0.00076
# edad_al_ing_1 33.351 1 7.7e-09
# ano_nac_corr 7.573 1 0.00592
# susinidum_oh 32.526 1 1.2e-08
# susinidum_coc 0.831 1 0.36209
# susinidum_pbc 2.136 1 0.14387
# susinidum_mar 22.932 1 1.7e-06
# psycom_dum_with 10.244 1 0.00137
# psycom_dum_study 41.095 1 1.4e-10
# freq_cons_dum_5day 15.029 1 0.00011
# cond_oc_dum_2inact 0.171 1 0.67918
# cond_oc_dum_3unemp 27.481 1 1.6e-07
# susprindum_oh 32.539 1 1.2e-08
# susprindum_coc 1.498 1 0.22104
# susprindum_pbc 30.540 1 3.3e-08
# susprindum_mar 0.970 1 0.32457
# GLOBAL 513.580 19 < 2e-16
message(paste0("There was no model that improved cox zph of the restricted sample (vs. ", round(cox.zph(cox_iiw_proposed)$table[nrow(cox.zph(cox_iiw_proposed)$table), 1],1),")"))Code
invisible("We improved fit from 552 to 513")
invisible("We do not longer try to change log_lag_dias_treat_imp_sin_na_rec")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("2024-04-30: restricted w/ tipo de plan 2 mod")
for (i in seq(.1,7,.1)) {
change_points <- c(5*i, 10*i, 15*i) # Example change points (months)
coefficients <- c(0.5, 1.5, 0.5, 1) # Example coefficients
# Add the time-varying coefficient to your model
data_mine_miss_restr_proc2$lag_less_90d_tr1_rec2 <- with(data_mine_miss_restr_proc2,
lag_less_90d_tr1*time_varying_coeff(lag_time, change_points, coefficients))
data_mine_miss_restr_proc2$less_90d_tr1_rec2 <- with(data_mine_miss_restr_proc2,
less_90d_tr1*time_varying_coeff(lag_time, change_points, coefficients))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec2+
lag_comp_bpsc_y3_severe +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d)
invisible("")
print(paste0("iteration: ",i))
print(change_points)
print(format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d)$table), 1]))
}[1] "iteration: 0.1"
[1] 0.5 1.0 1.5
[1] "493.4811"
[1] "iteration: 0.2"
[1] 1 2 3
[1] "493.2048"
[1] "iteration: 0.3"
[1] 1.5 3.0 4.5
[1] "492.4265"
[1] "iteration: 0.4"
[1] 2 4 6
[1] "480.5242"
[1] "iteration: 0.5"
[1] 2.5 5.0 7.5
[1] "477.4635"
[1] "iteration: 0.6"
[1] 3 6 9
[1] "482.7047"
[1] "iteration: 0.7"
[1] 3.5 7.0 10.5
[1] "482.1585"
[1] "iteration: 0.8"
[1] 4 8 12
[1] "479.9969"
[1] "iteration: 0.9"
[1] 4.5 9.0 13.5
[1] "477.7753"
[1] "iteration: 1"
[1] 5 10 15
[1] "479.6782"
[1] "iteration: 1.1"
[1] 5.5 11.0 16.5
[1] "480.8076"
[1] "iteration: 1.2"
[1] 6 12 18
[1] "478.0329"
[1] "iteration: 1.3"
[1] 6.5 13.0 19.5
[1] "477.8009"
[1] "iteration: 1.4"
[1] 7 14 21
[1] "478.0451"
[1] "iteration: 1.5"
[1] 7.5 15.0 22.5
[1] "477.306"
[1] "iteration: 1.6"
[1] 8 16 24
[1] "476.7152"
[1] "iteration: 1.7"
[1] 8.5 17.0 25.5
[1] "475.7639"
[1] "iteration: 1.8"
[1] 9 18 27
[1] "477.6404"
[1] "iteration: 1.9"
[1] 9.5 19.0 28.5
[1] "478.1175"
[1] "iteration: 2"
[1] 10 20 30
[1] "478.6989"
[1] "iteration: 2.1"
[1] 10.5 21.0 31.5
[1] "477.3389"
[1] "iteration: 2.2"
[1] 11 22 33
[1] "477.742"
[1] "iteration: 2.3"
[1] 11.5 23.0 34.5
[1] "479.8333"
[1] "iteration: 2.4"
[1] 12 24 36
[1] "480.6214"
[1] "iteration: 2.5"
[1] 12.5 25.0 37.5
[1] "479.5284"
[1] "iteration: 2.6"
[1] 13 26 39
[1] "479.3746"
[1] "iteration: 2.7"
[1] 13.5 27.0 40.5
[1] "479.5649"
[1] "iteration: 2.8"
[1] 14 28 42
[1] "480.0663"
[1] "iteration: 2.9"
[1] 14.5 29.0 43.5
[1] "482.4381"
[1] "iteration: 3"
[1] 15 30 45
[1] "483.276"
[1] "iteration: 3.1"
[1] 15.5 31.0 46.5
[1] "485.369"
[1] "iteration: 3.2"
[1] 16 32 48
[1] "488.6019"
[1] "iteration: 3.3"
[1] 16.5 33.0 49.5
[1] "488.7055"
[1] "iteration: 3.4"
[1] 17 34 51
[1] "490.8723"
[1] "iteration: 3.5"
[1] 17.5 35.0 52.5
[1] "490.1774"
[1] "iteration: 3.6"
[1] 18 36 54
[1] "491.0848"
[1] "iteration: 3.7"
[1] 18.5 37.0 55.5
[1] "491.8408"
[1] "iteration: 3.8"
[1] 19 38 57
[1] "493.2258"
[1] "iteration: 3.9"
[1] 19.5 39.0 58.5
[1] "493.0619"
[1] "iteration: 4"
[1] 20 40 60
[1] "494.1005"
[1] "iteration: 4.1"
[1] 20.5 41.0 61.5
[1] "495.7924"
[1] "iteration: 4.2"
[1] 21 42 63
[1] "496.9032"
[1] "iteration: 4.3"
[1] 21.5 43.0 64.5
[1] "497.0522"
[1] "iteration: 4.4"
[1] 22 44 66
[1] "497.3046"
[1] "iteration: 4.5"
[1] 22.5 45.0 67.5
[1] "497.4102"
[1] "iteration: 4.6"
[1] 23 46 69
[1] "497.9092"
[1] "iteration: 4.7"
[1] 23.5 47.0 70.5
[1] "498.3643"
[1] "iteration: 4.8"
[1] 24 48 72
[1] "498.6149"
[1] "iteration: 4.9"
[1] 24.5 49.0 73.5
[1] "500.2164"
[1] "iteration: 5"
[1] 25 50 75
[1] "500.4883"
[1] "iteration: 5.1"
[1] 25.5 51.0 76.5
[1] "501.8345"
[1] "iteration: 5.2"
[1] 26 52 78
[1] "502.9337"
[1] "iteration: 5.3"
[1] 26.5 53.0 79.5
[1] "503.2239"
[1] "iteration: 5.4"
[1] 27 54 81
[1] "502.2895"
[1] "iteration: 5.5"
[1] 27.5 55.0 82.5
[1] "502.7881"
[1] "iteration: 5.6"
[1] 28 56 84
[1] "503.4912"
[1] "iteration: 5.7"
[1] 28.5 57.0 85.5
[1] "503.9953"
[1] "iteration: 5.8"
[1] 29 58 87
[1] "504.2127"
[1] "iteration: 5.9"
[1] 29.5 59.0 88.5
[1] "504.5094"
[1] "iteration: 6"
[1] 30 60 90
[1] "504.3318"
[1] "iteration: 6.1"
[1] 30.5 61.0 91.5
[1] "503.4482"
[1] "iteration: 6.2"
[1] 31 62 93
[1] "503.827"
[1] "iteration: 6.3"
[1] 31.5 63.0 94.5
[1] "503.1115"
[1] "iteration: 6.4"
[1] 32 64 96
[1] "503.6159"
[1] "iteration: 6.5"
[1] 32.5 65.0 97.5
[1] "503.661"
[1] "iteration: 6.6"
[1] 33 66 99
[1] "503.5745"
[1] "iteration: 6.7"
[1] 33.5 67.0 100.5
[1] "503.7049"
[1] "iteration: 6.8"
[1] 34 68 102
[1] "503.1642"
[1] "iteration: 6.9"
[1] 34.5 69.0 103.5
[1] "503.5052"
[1] "iteration: 7"
[1] 35 70 105
[1] "502.6516"Code
# [1] "iteration: 0.5"
# [1] 2.5 5.0 7.5
# [1] "475.9965"
message(paste0("There was no model that improved cox zph of the restricted sample (vs. ", round(cox.zph(cox_iiw_proposed)$table[nrow(cox.zph(cox_iiw_proposed)$table), 1],1),")"))Code
#There was no model that improved cox zph of the restricted sample (vs. 495.5)
invisible("Grid search of reduction of TVCs")
for (i in seq(.1,1.5,.05)) {
tvcs <- c(20,80)*i
data_mine_miss_restr_proc2$lag_less_90d_tr1_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$lag_less_90d_tr1,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
data_mine_miss_restr_proc2$less_90d_tr1_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$less_90d_tr1,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt)
invisible("")
print(paste0("iteration: ",i))
print(tvcs)
print(format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt)$table), 1]))
}[1] "iteration: 0.1"
[1] 2 8
[1] "1710.118"
[1] "iteration: 0.15"
[1] 3 12
[1] "1642.617"
[1] "iteration: 0.2"
[1] 4 16
[1] "1319.002"
[1] "iteration: 0.25"
[1] 5 20
[1] "885.2445"
[1] "iteration: 0.3"
[1] 6 24
[1] "503.2978"
[1] "iteration: 0.35"
[1] 7 28
[1] "363.183"
[1] "iteration: 0.4"
[1] 8 32
[1] "609.1047"
[1] "iteration: 0.45"
[1] 9 36
[1] "1241.478"
[1] "iteration: 0.5"
[1] 10 40
[1] "2116.105"
[1] "iteration: 0.55"
[1] 11 44
[1] "3014.795"
[1] "iteration: 0.6"
[1] 12 48
[1] "3752.182"
[1] "iteration: 0.65"
[1] 13 52
[1] "4241.554"
[1] "iteration: 0.7"
[1] 14 56
[1] "4487.798"
[1] "iteration: 0.75"
[1] 15 60
[1] "4534.477"
[1] "iteration: 0.8"
[1] 16 64
[1] "4426.349"
[1] "iteration: 0.85"
[1] 17 68
[1] "4208.187"
[1] "iteration: 0.9"
[1] 18 72
[1] "3908.978"
[1] "iteration: 0.95"
[1] 19 76
[1] "3553.854"
[1] "iteration: 1"
[1] 20 80
[1] "3162.509"
[1] "iteration: 1.05"
[1] 21 84
[1] "2756.184"
[1] "iteration: 1.1"
[1] 22 88
[1] "2339.299"
[1] "iteration: 1.15"
[1] 23 92
[1] "1922.443"
[1] "iteration: 1.2"
[1] 24 96
[1] "1526.312"
[1] "iteration: 1.25"
[1] 25 100
[1] "1186.425"
[1] "iteration: 1.3"
[1] 26 104
[1] "943.2429"
[1] "iteration: 1.35"
[1] 27 108
[1] "826.6724"
[1] "iteration: 1.4"
[1] 28 112
[1] "871.2033"
[1] "iteration: 1.45"
[1] 29 116
[1] "1124.667"
[1] "iteration: 1.5"
[1] 30 120
[1] "1270.214"Code
# [1] "iteration: 0.35"
# [1] 7 28
# [1] "361.9089"
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
# [1] "iteration: 0.5"
# [1] 2.5 5.0 7.5
# [1] "475.9965"
change_points <- c(2.5, 5, 7.5) # Example change points (months)
coefficients <- c(0.5, 1.5, 0.5, 1) # Example coefficients
# Add the time-varying coefficient to your model
data_mine_miss_restr_proc2$lag_less_90d_tr1_rec2 <- with(data_mine_miss_restr_proc2,
lag_less_90d_tr1*time_varying_coeff(lag_time, change_points, coefficients))
data_mine_miss_restr_proc2$less_90d_tr1_rec2 <- with(data_mine_miss_restr_proc2,
less_90d_tr1*time_varying_coeff(lag_time, change_points, coefficients))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d<-
cph(Surv(lag_time,time,event)~
cluster(id) +
lag_tr_outcome +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec2 +
lag_comp_bpsc_y3_severe +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod),
#data= data_clean,
data= data_mine_miss_restr_proc2 %>% dplyr::mutate(log_lag_dias_treat_imp_sin_na_rec= bs(log_lag_dias_treat_imp_sin_na, 3)) %>% data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d) chisq df p
lag_tr_outcome 54.041 1 2.0e-13
log_lag_dias_treat_imp_sin_na 144.161 1 < 2e-16
lag_less_90d_tr1_rec2 53.529 1 2.5e-13
lag_comp_bpsc_y3_severe 21.026 1 4.5e-06
lag_policonsumo2 8.345 1 0.00387
edad_al_ing_1 33.844 1 6.0e-09
ano_nac_corr 8.075 1 0.00449
susinidum_oh 40.358 1 2.1e-10
susinidum_coc 0.976 1 0.32311
susinidum_pbc 6.830 1 0.00897
susinidum_mar 25.019 1 5.7e-07
psycom_dum_with 17.112 1 3.5e-05
psycom_dum_study 40.295 1 2.2e-10
freq_cons_dum_5day 2.353 1 0.12505
cond_oc_dum_2inact 3.057 1 0.08040
cond_oc_dum_3unemp 14.775 1 0.00012
susprindum_oh 25.139 1 5.3e-07
susprindum_coc 0.780 1 0.37706
susprindum_pbc 20.331 1 6.5e-06
susprindum_mar 2.509 1 0.11319
GLOBAL 477.463 20 < 2e-16Code
message(paste0("The model improved from ", round(cox.zph(cox_iiw_proposed)$table[nrow(cox.zph(cox_iiw_proposed)$table), 1],1)," to 476"))Code
#The model improved from 495.5 to 476
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
# [1] "iteration: 0.35"
# [1] 7 28
# [1] "361.9089"
tvcs <- c(20,80)*0.35
data_mine_miss_restr_proc2$lag_less_90d_tr1_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$lag_less_90d_tr1,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
data_mine_miss_restr_proc2$less_90d_tr1_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$less_90d_tr1,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt) chisq df p
lag_tr_outcome 20.783 1 5.1e-06
log_lag_dias_treat_imp_sin_na 41.447 1 1.2e-10
lag_less_90d_tr1_rec 14.941 1 0.00011
lag_comp_bpsc_y3_severe 16.927 1 3.9e-05
lag_policonsumo2 8.209 1 0.00417
edad_al_ing_1 20.696 1 5.4e-06
ano_nac_corr 2.130 1 0.14448
susinidum_oh 33.052 1 9.0e-09
susinidum_coc 1.324 1 0.24985
susinidum_pbc 3.595 1 0.05794
susinidum_mar 22.314 1 2.3e-06
psycom_dum_with 6.947 1 0.00840
psycom_dum_study 11.181 1 0.00083
freq_cons_dum_5day 1.635 1 0.20096
cond_oc_dum_2inact 3.157 1 0.07561
cond_oc_dum_3unemp 12.212 1 0.00047
susprindum_oh 19.510 1 1.0e-05
susprindum_coc 0.602 1 0.43788
susprindum_pbc 13.440 1 0.00025
susprindum_mar 4.138 1 0.04192
GLOBAL 363.183 20 < 2e-16Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Analyse the violation of proportional hazards")
par(mar = c(5, 4, 4, 5))
par(mfrow=c(3,1))
res240422_2 <- resid(model_after_mod_not_log_lag_dias_treat_lag_less_90d, "scaledsch")
time240422_2 <- as.numeric(dimnames(res240422_2)[[1]])
z_240422_2_1 <- loess(res240422_2[,"log_lag_dias_treat_imp_sin_na"] ~ time240422_2, span=0.50) # residuals
# x11()
plot(time240422_2, fitted(z_240422_2_1))
lines(supsmu(time240422_2, res240422_2[,"log_lag_dias_treat_imp_sin_na"]),lty=2)
title("Sch. res., log_lag_dias_treat_imp_sin_na")
z_240422_2_2 <- loess(res240422_2[,"lag_comp_bpsc_y3_severe"] ~ time240422_2, span=0.50) #
plot(time240422_2, fitted(z_240422_2_2))
lines(supsmu(time240422_2, res240422_2[,"lag_comp_bpsc_y3_severe"]),lty=2)
title("Sch. res., lag_comp_bpsc_y3_severe")
km_lag_comp_bpsc_y3_severe <- survfit(Surv(lag_time,time,event)~ lag_comp_bpsc_y3_severe+
cluster(id), data = data_mine_miss_proc2)
#autoplot(km_home) # just to see the km curves
plot(km_lag_comp_bpsc_y3_severe, fun = "cloglog", xlab = "Time (in days) using log",col=c("black", "red"),
ylab = "log-log survival", main = "log-log curves by km_lag_comp_bpsc_y3_severe")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1)
Code
recorded_plot <- recordPlot()
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
png(paste0(folder_path,"zph/proportionallity_comp_bpsc.png"), height=13, width=10, res=500, units="in")
recorded_plot
dev.off()png
2 Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("2024-04-30: Analyse the violation of proportional hazards (restricted)")
par(mar = c(5, 4, 4, 5))
par(mfrow=c(3,1))
res240422_2_restr <- resid(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt, "scaledsch")
time240422_2_restr <- as.numeric(dimnames(res240422_2_restr)[[1]])
z_240422_2_1_restr <- loess(res240422_2_restr[,"log_lag_dias_treat_imp_sin_na"] ~ time240422_2_restr, span=0.50) # residuals
# x11()
plot(time240422_2_restr, fitted(z_240422_2_1_restr))
lines(supsmu(time240422_2_restr, res240422_2_restr[,"log_lag_dias_treat_imp_sin_na"]),lty=2)
title("Sch. res., log_lag_dias_treat_imp_sin_na (restricted)")
z_240422_2_2_restr <- loess(res240422_2_restr[,"lag_comp_bpsc_y3_severe"] ~ time240422_2_restr, span=0.50) #
plot(time240422_2_restr, fitted(z_240422_2_2_restr))
lines(supsmu(time240422_2_restr, res240422_2_restr[,"lag_comp_bpsc_y3_severe"]),lty=2)
title("Sch. res., lag_comp_bpsc_y3_severe (restricted)")
km_lag_comp_bpsc_y3_severe_restr <- survfit(Surv(lag_time,time,event)~ lag_comp_bpsc_y3_severe+
cluster(id), data = data_mine_miss_restr_proc2)
#autoplot(km_home) # just to see the km curves
plot(km_lag_comp_bpsc_y3_severe_restr, fun = "cloglog", xlab = "Time (in days) using log",col=c("black", "red"),
ylab = "log-log survival", main = "log-log curves by km_lag_comp_bpsc_y3_severe")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1)
Code
recorded_plot <- recordPlot()
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
png(paste0(folder_path,"zph/proportionallity_comp_bpsc_restr.png"), height=13, width=10, res=500, units="in")
recorded_plot
dev.off()png
2 Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("2024-05-30: restricted w/ tipo de plan 2 mod; find TVCs")
message(paste0("The model was better after correction: ",
round(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt)$table), 1],1)
," vs. the original: ",
round(cox.zph(cox_iiw_proposed)$table[nrow(cox.zph(cox_iiw_proposed)$table), 1],1)
))Code
#The model was better after correction: 361.9 vs. the original: 495.5
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
for (i in seq(.1,7,.1)) {
change_points <- c(5*i, 10*i, 15*i) # Example change points (months)
coefficients <- c(0.5, 1.5, 0.5, 1) # Example coefficients
# Add the time-varying coefficient to your model
data_mine_miss_restr_proc2$lag_comp_bpsc_y3_severe_rec2 <- with(data_mine_miss_restr_proc2, lag_comp_bpsc_y3_severe*time_varying_coeff(lag_time, change_points, coefficients))
data_mine_miss_restr_proc2$comp_bpsc_y3_severe_rec2 <- with(data_mine_miss_restr_proc2, comp_bpsc_y3_severe*time_varying_coeff(time, change_points, coefficients))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec2 +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe)
invisible("")
print(paste0("iteration: ",i))
print(change_points)
print(format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe)$table), 1]))
}[1] "iteration: 0.1"
[1] 0.5 1.0 1.5
[1] "366.7352"
[1] "iteration: 0.2"
[1] 1 2 3
[1] "367.6181"
[1] "iteration: 0.3"
[1] 1.5 3.0 4.5
[1] "359.0755"
[1] "iteration: 0.4"
[1] 2 4 6
[1] "358.9647"
[1] "iteration: 0.5"
[1] 2.5 5.0 7.5
[1] "362.0432"
[1] "iteration: 0.6"
[1] 3 6 9
[1] "367.2566"
[1] "iteration: 0.7"
[1] 3.5 7.0 10.5
[1] "368.7599"
[1] "iteration: 0.8"
[1] 4 8 12
[1] "364.0977"
[1] "iteration: 0.9"
[1] 4.5 9.0 13.5
[1] "363.4215"
[1] "iteration: 1"
[1] 5 10 15
[1] "361.3236"
[1] "iteration: 1.1"
[1] 5.5 11.0 16.5
[1] "361.5829"
[1] "iteration: 1.2"
[1] 6 12 18
[1] "359.3325"
[1] "iteration: 1.3"
[1] 6.5 13.0 19.5
[1] "357.4246"
[1] "iteration: 1.4"
[1] 7 14 21
[1] "356.7213"
[1] "iteration: 1.5"
[1] 7.5 15.0 22.5
[1] "356.5751"
[1] "iteration: 1.6"
[1] 8 16 24
[1] "356.1985"
[1] "iteration: 1.7"
[1] 8.5 17.0 25.5
[1] "355.9476"
[1] "iteration: 1.8"
[1] 9 18 27
[1] "356.0388"
[1] "iteration: 1.9"
[1] 9.5 19.0 28.5
[1] "355.7839"
[1] "iteration: 2"
[1] 10 20 30
[1] "355.8856"
[1] "iteration: 2.1"
[1] 10.5 21.0 31.5
[1] "355.274"
[1] "iteration: 2.2"
[1] 11 22 33
[1] "355.3662"
[1] "iteration: 2.3"
[1] 11.5 23.0 34.5
[1] "355.2717"
[1] "iteration: 2.4"
[1] 12 24 36
[1] "355.2632"
[1] "iteration: 2.5"
[1] 12.5 25.0 37.5
[1] "355.6316"
[1] "iteration: 2.6"
[1] 13 26 39
[1] "355.7911"
[1] "iteration: 2.7"
[1] 13.5 27.0 40.5
[1] "355.9782"
[1] "iteration: 2.8"
[1] 14 28 42
[1] "355.725"
[1] "iteration: 2.9"
[1] 14.5 29.0 43.5
[1] "356.2413"
[1] "iteration: 3"
[1] 15 30 45
[1] "356.405"
[1] "iteration: 3.1"
[1] 15.5 31.0 46.5
[1] "356.2095"
[1] "iteration: 3.2"
[1] 16 32 48
[1] "357.5944"
[1] "iteration: 3.3"
[1] 16.5 33.0 49.5
[1] "358.5072"
[1] "iteration: 3.4"
[1] 17 34 51
[1] "359.3226"
[1] "iteration: 3.5"
[1] 17.5 35.0 52.5
[1] "359.6748"
[1] "iteration: 3.6"
[1] 18 36 54
[1] "359.214"
[1] "iteration: 3.7"
[1] 18.5 37.0 55.5
[1] "359.4672"
[1] "iteration: 3.8"
[1] 19 38 57
[1] "361.2473"
[1] "iteration: 3.9"
[1] 19.5 39.0 58.5
[1] "361.1348"
[1] "iteration: 4"
[1] 20 40 60
[1] "361.7292"
[1] "iteration: 4.1"
[1] 20.5 41.0 61.5
[1] "364.4573"
[1] "iteration: 4.2"
[1] 21 42 63
[1] "363.8848"
[1] "iteration: 4.3"
[1] 21.5 43.0 64.5
[1] "363.5723"
[1] "iteration: 4.4"
[1] 22 44 66
[1] "364.5945"
[1] "iteration: 4.5"
[1] 22.5 45.0 67.5
[1] "365.1071"
[1] "iteration: 4.6"
[1] 23 46 69
[1] "365.8701"
[1] "iteration: 4.7"
[1] 23.5 47.0 70.5
[1] "364.955"
[1] "iteration: 4.8"
[1] 24 48 72
[1] "365.3112"
[1] "iteration: 4.9"
[1] 24.5 49.0 73.5
[1] "365.706"
[1] "iteration: 5"
[1] 25 50 75
[1] "366.3371"
[1] "iteration: 5.1"
[1] 25.5 51.0 76.5
[1] "367.0493"
[1] "iteration: 5.2"
[1] 26 52 78
[1] "367.9306"
[1] "iteration: 5.3"
[1] 26.5 53.0 79.5
[1] "368.4692"
[1] "iteration: 5.4"
[1] 27 54 81
[1] "368.9484"
[1] "iteration: 5.5"
[1] 27.5 55.0 82.5
[1] "370.2654"
[1] "iteration: 5.6"
[1] 28 56 84
[1] "370.4634"
[1] "iteration: 5.7"
[1] 28.5 57.0 85.5
[1] "371.0471"
[1] "iteration: 5.8"
[1] 29 58 87
[1] "371.8349"
[1] "iteration: 5.9"
[1] 29.5 59.0 88.5
[1] "372.2448"
[1] "iteration: 6"
[1] 30 60 90
[1] "371.6"
[1] "iteration: 6.1"
[1] 30.5 61.0 91.5
[1] "371.2923"
[1] "iteration: 6.2"
[1] 31 62 93
[1] "371.9248"
[1] "iteration: 6.3"
[1] 31.5 63.0 94.5
[1] "372.2759"
[1] "iteration: 6.4"
[1] 32 64 96
[1] "371.9611"
[1] "iteration: 6.5"
[1] 32.5 65.0 97.5
[1] "371.9074"
[1] "iteration: 6.6"
[1] 33 66 99
[1] "371.3031"
[1] "iteration: 6.7"
[1] 33.5 67.0 100.5
[1] "371.2377"
[1] "iteration: 6.8"
[1] 34 68 102
[1] "370.6362"
[1] "iteration: 6.9"
[1] 34.5 69.0 103.5
[1] "371.8994"
[1] "iteration: 7"
[1] 35 70 105
[1] "371.8839"Code
message(paste0("The previously modified model: ",
round(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt)$table), 1],1),
" vs. the actual models: 354.7"
))Code
#The previously modified model: 361.9 vs. the actual models: 354.7
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_bpsc_severe <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq(.1,1.5,.025)) {
tvcs <- c(20,80)*i
data_mine_miss_restr_proc2$lag_comp_bpsc_y3_severe_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$lag_comp_bpsc_y3_severe,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
data_mine_miss_restr_proc2$comp_bpsc_y3_severe_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$comp_bpsc_y3_severe,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt)
invisible("")
print(paste0("iteration: ",i))
aic_iter <- round(extractAIC(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt)[[2]],1)
print( paste0("Time points: ",tvcs[1],", ", tvcs[2]))
print( paste0("AIC: ", aic_iter))
cox_zph_iter <- format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt)$table), 1])
print(paste0("Cox ZPH: ", cox_zph_iter))
results_bpsc_severe <- rbind.data.frame(results_bpsc_severe,
cbind.data.frame(
spec = paste0("iteration: ",i, "; tvcs: ", tvcs[1],", ", tvcs[2]),
aic = aic_iter,
cox_zph = cox_zph_iter,
success = T
))
}[1] "iteration: 0.1"
[1] "Time points: 2, 8"
[1] "AIC: 244648.5"
[1] "Cox ZPH: 2027.052"
[1] "iteration: 0.125"
[1] "Time points: 2.5, 10"
[1] "AIC: 244584.1"
[1] "Cox ZPH: 2126.697"
[1] "iteration: 0.15"
[1] "Time points: 3, 12"
[1] "AIC: 244529.8"
[1] "Cox ZPH: 2176.088"
[1] "iteration: 0.175"
[1] "Time points: 3.5, 14"
[1] "AIC: 244488.4"
[1] "Cox ZPH: 2165.634"
[1] "iteration: 0.2"
[1] "Time points: 4, 16"
[1] "AIC: 244463.4"
[1] "Cox ZPH: 2084.661"
[1] "iteration: 0.225"
[1] "Time points: 4.5, 18"
[1] "AIC: 244450.7"
[1] "Cox ZPH: 1955.693"
[1] "iteration: 0.25"
[1] "Time points: 5, 20"
[1] "AIC: 244448.3"
[1] "Cox ZPH: 1792.717"
[1] "iteration: 0.275"
[1] "Time points: 5.5, 22"
[1] "AIC: 244464.3"
[1] "Cox ZPH: 1576.648"
[1] "iteration: 0.3"
[1] "Time points: 6, 24"
[1] "AIC: 244495.7"
[1] "Cox ZPH: 1325.06"
[1] "iteration: 0.325"
[1] "Time points: 6.5, 26"
[1] "AIC: 244537.3"
[1] "Cox ZPH: 1069.782"
[1] "iteration: 0.35"
[1] "Time points: 7, 28"
[1] "AIC: 244590.8"
[1] "Cox ZPH: 825.2673"
[1] "iteration: 0.375"
[1] "Time points: 7.5, 30"
[1] "AIC: 244656"
[1] "Cox ZPH: 608.8866"
[1] "iteration: 0.4"
[1] "Time points: 8, 32"
[1] "AIC: 244730.5"
[1] "Cox ZPH: 444.7581"
[1] "iteration: 0.425"
[1] "Time points: 8.5, 34"
[1] "AIC: 244815.5"
[1] "Cox ZPH: 352.1245"
[1] "iteration: 0.45"
[1] "Time points: 9, 36"
[1] "AIC: 244909.7"
[1] "Cox ZPH: 353.8899"
[1] "iteration: 0.475"
[1] "Time points: 9.5, 38"
[1] "AIC: 245009.8"
[1] "Cox ZPH: 465.3513"
[1] "iteration: 0.5"
[1] "Time points: 10, 40"
[1] "AIC: 245111.2"
[1] "Cox ZPH: 687.6218"
[1] "iteration: 0.525"
[1] "Time points: 10.5, 42"
[1] "AIC: 245215"
[1] "Cox ZPH: 1018.352"
[1] "iteration: 0.55"
[1] "Time points: 11, 44"
[1] "AIC: 245321.5"
[1] "Cox ZPH: 1450.489"
[1] "iteration: 0.575"
[1] "Time points: 11.5, 46"
[1] "AIC: 245426.3"
[1] "Cox ZPH: 1956.628"
[1] "iteration: 0.6"
[1] "Time points: 12, 48"
[1] "AIC: 245526.3"
[1] "Cox ZPH: 2500.774"
[1] "iteration: 0.625"
[1] "Time points: 12.5, 50"
[1] "AIC: 245620.4"
[1] "Cox ZPH: 3058.768"
[1] "iteration: 0.65"
[1] "Time points: 13, 52"
[1] "AIC: 245707.3"
[1] "Cox ZPH: 3601.197"
[1] "iteration: 0.675"
[1] "Time points: 13.5, 54"
[1] "AIC: 245785.6"
[1] "Cox ZPH: 4103.462"
[1] "iteration: 0.7"
[1] "Time points: 14, 56"
[1] "AIC: 245853"
[1] "Cox ZPH: 4540.246"
[1] "iteration: 0.725"
[1] "Time points: 14.5, 58"
[1] "AIC: 245907.2"
[1] "Cox ZPH: 4898.792"
[1] "iteration: 0.75"
[1] "Time points: 15, 60"
[1] "AIC: 245946"
[1] "Cox ZPH: 5174.502"
[1] "iteration: 0.775"
[1] "Time points: 15.5, 62"
[1] "AIC: 245968.4"
[1] "Cox ZPH: 5365.95"
[1] "iteration: 0.8"
[1] "Time points: 16, 64"
[1] "AIC: 245972.6"
[1] "Cox ZPH: 5481.321"
[1] "iteration: 0.825"
[1] "Time points: 16.5, 66"
[1] "AIC: 245957.5"
[1] "Cox ZPH: 5527.575"
[1] "iteration: 0.85"
[1] "Time points: 17, 68"
[1] "AIC: 245922.4"
[1] "Cox ZPH: 5514.13"
[1] "iteration: 0.875"
[1] "Time points: 17.5, 70"
[1] "AIC: 245866.7"
[1] "Cox ZPH: 5451.26"
[1] "iteration: 0.9"
[1] "Time points: 18, 72"
[1] "AIC: 245789.7"
[1] "Cox ZPH: 5347.218"
[1] "iteration: 0.925"
[1] "Time points: 18.5, 74"
[1] "AIC: 245690.6"
[1] "Cox ZPH: 5203.803"
[1] "iteration: 0.95"
[1] "Time points: 19, 76"
[1] "AIC: 245569.6"
[1] "Cox ZPH: 5028.824"
[1] "iteration: 0.975"
[1] "Time points: 19.5, 78"
[1] "AIC: 245425.6"
[1] "Cox ZPH: 4827.561"
[1] "iteration: 1"
[1] "Time points: 20, 80"
[1] "AIC: 245258.7"
[1] "Cox ZPH: 4603.399"
[1] "iteration: 1.025"
[1] "Time points: 20.5, 82"
[1] "AIC: 245070.3"
[1] "Cox ZPH: 4363.59"
[1] "iteration: 1.05"
[1] "Time points: 21, 84"
[1] "AIC: 244861.8"
[1] "Cox ZPH: 4111.058"
[1] "iteration: 1.075"
[1] "Time points: 21.5, 86"
[1] "AIC: 244633.9"
[1] "Cox ZPH: 3848.298"
[1] "iteration: 1.1"
[1] "Time points: 22, 88"
[1] "AIC: 244389.1"
[1] "Cox ZPH: 3579.744"
[1] "iteration: 1.125"
[1] "Time points: 22.5, 90"
[1] "AIC: 244129.8"
[1] "Cox ZPH: 3308.287"
[1] "iteration: 1.15"
[1] "Time points: 23, 92"
[1] "AIC: 243857.6"
[1] "Cox ZPH: 3037.355"
[1] "iteration: 1.175"
[1] "Time points: 23.5, 94"
[1] "AIC: 243578.2"
[1] "Cox ZPH: 2772.307"
[1] "iteration: 1.2"
[1] "Time points: 24, 96"
[1] "AIC: 243291.8"
[1] "Cox ZPH: 2512.812"
[1] "iteration: 1.225"
[1] "Time points: 24.5, 98"
[1] "AIC: 243000.5"
[1] "Cox ZPH: 2265.902"
[1] "iteration: 1.25"
[1] "Time points: 25, 100"
[1] "AIC: 242705.8"
[1] "Cox ZPH: 2025.213"
[1] "iteration: 1.275"
[1] "Time points: 25.5, 102"
[1] "AIC: 242407"
[1] "Cox ZPH: 1795.87"
[1] "iteration: 1.3"
[1] "Time points: 26, 104"
[1] "AIC: 242104.5"
[1] "Cox ZPH: 1579.576"
[1] "iteration: 1.325"
[1] "Time points: 26.5, 106"
[1] "AIC: 241802.2"
[1] "Cox ZPH: 1380.713"
[1] "iteration: 1.35"
[1] "Time points: 27, 108"
[1] "AIC: 241506.2"
[1] "Cox ZPH: 1203.313"
[1] "iteration: 1.375"
[1] "Time points: 27.5, 110"
[1] "AIC: 241224.7"
[1] "Cox ZPH: 1051.053"
[1] "iteration: 1.4"
[1] "Time points: 28, 112"
[1] "AIC: 240963.6"
[1] "Cox ZPH: 927.1896"
[1] "iteration: 1.425"
[1] "Time points: 28.5, 114"
[1] "AIC: 240715"
[1] "Cox ZPH: 832.7559"
[1] "iteration: 1.45"
[1] "Time points: 29, 116"
[1] "AIC: 240498.4"
[1] "Cox ZPH: 775.9883"
[1] "iteration: 1.475"
[1] "Time points: 29.5, 118"
[1] "AIC: 240307.3"
[1] "Cox ZPH: 769.6182"
[1] "iteration: 1.5"
[1] "Time points: 30, 120"
[1] "AIC: 240155.7"
[1] "Cox ZPH: 837.3566"Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_bpsc_severe[which(as.numeric(results_bpsc_severe$cox_zph)==
min(as.numeric(results_bpsc_severe$cox_zph))),] spec aic cox_zph success
14 iteration: 0.425; tvcs: 8.5, 34 244815.5 352.1245 TRUECode
# spec aic cox_zph success
# 14 iteration: 0.425; tvcs: 8.5, 34 244802.6 351.5909 TRUE
tvcs <- c(20,80)*0.425
data_mine_miss_restr_proc2$lag_comp_bpsc_y3_severe_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$lag_comp_bpsc_y3_severe,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
data_mine_miss_restr_proc2$comp_bpsc_y3_severe_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$comp_bpsc_y3_severe,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt<-
cph(Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt) chisq df p
lag_tr_outcome 19.1955 1 1.2e-05
log_lag_dias_treat_imp_sin_na 49.0939 1 2.4e-12
lag_less_90d_tr1_rec 42.0858 1 8.7e-11
lag_comp_bpsc_y3_severe_rec 1.5266 1 0.21662
lag_policonsumo2 1.5932 1 0.20687
edad_al_ing_1 17.6296 1 2.7e-05
ano_nac_corr 1.6974 1 0.19263
susinidum_oh 23.7232 1 1.1e-06
susinidum_coc 3.6139 1 0.05730
susinidum_pbc 2.7404 1 0.09784
susinidum_mar 13.2315 1 0.00028
psycom_dum_with 14.6452 1 0.00013
psycom_dum_study 8.0507 1 0.00455
freq_cons_dum_5day 0.4958 1 0.48136
cond_oc_dum_2inact 7.9362 1 0.00485
cond_oc_dum_3unemp 5.2995 1 0.02133
susprindum_oh 11.5928 1 0.00066
susprindum_coc 0.0487 1 0.82530
susprindum_pbc 2.8441 1 0.09171
susprindum_mar 7.8036 1 0.00521
GLOBAL 352.1245 20 < 2e-16Code
# chisq df p
# lag_tr_outcome 19.4440 1 1.0e-05 ###
# log_lag_dias_treat_imp_sin_na 48.7236 1 2.9e-12 ####
# lag_less_90d_tr1_rec 42.0653 1 8.8e-11 ####
# lag_comp_bpsc_y3_severe_rec 1.5369 1 0.21508
# lag_policonsumo2 1.5765 1 0.20926
# edad_al_ing_1 17.5785 1 2.8e-05 ###
# ano_nac_corr 1.6876 1 0.19392
# susinidum_oh 23.5183 1 1.2e-06 ###
# susinidum_coc 3.6019 1 0.05771
# susinidum_pbc 2.6388 1 0.10428
# susinidum_mar 13.2146 1 0.00028 ##
# psycom_dum_with 14.7208 1 0.00012 #
# psycom_dum_study 8.0570 1 0.00453 #
# freq_cons_dum_5day 0.5748 1 0.44834
# cond_oc_dum_2inact 7.9683 1 0.00476
# cond_oc_dum_3unemp 5.0871 1 0.02410 #
# susprindum_oh 11.4536 1 0.00071 ##
# susprindum_coc 0.0549 1 0.81473
# susprindum_pbc 2.7348 1 0.09818
# susprindum_mar 7.8675 1 0.00503 ##
# GLOBAL 351.5909 20 < 2e-16
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Analyse the violation of proportional hazards")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("2024-04-30: restricted")
res240422_3_restr <- resid(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt, "scaledsch")
time240422_3_restr <- as.numeric(dimnames(res240422_3_restr)[[1]])
z_240422_3_1_restr <- loess(res240422_3_restr[,"susinidum_oh"] ~ time240422_3_restr, span=0.50) # residuals
# x11()
par(mfrow= c(3,2))
plot(time240422_3_restr, fitted(z_240422_3_1_restr))
lines(supsmu(time240422_3_restr, res240422_3_restr[,"susinidum_oh"]),lty=2)
title("Sch. res., susinidum_oh")
z_240422_3_2_restr <- loess(res240422_3_restr[,"susinidum_oh"] ~ time240422_3_restr, span=0.50) #
plot(time240422_3_restr, fitted(z_240422_3_2_restr))
lines(supsmu(time240422_3_restr, res240422_3_restr[,"susinidum_oh"]),lty=2)
title("Sch. res., susinidum_oh")
plot(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt)[which(attr(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt$coefficients, "names") == "psycom_dum_study")], lwd=2, main= "Residuals of psycom_dum_study")
abline(0,0, col="red", lty=3, lwd=2)
abline(h=model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt$coefficients["psycom_dum_study"], col=3, lwd=2, lty= 3)
legend("bottomleft", legend= c("Reference line for the null effect", "Average hazard over time", "Time-varying hazard"), lty=c(3,2,1), col= c("red",3,1), lwd=2, bty = "n", bg = "transparent")
plot(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt)[which(attr(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d$coefficients, "names") == "psycom_dum_with")], lwd=2, main= "Residuals of psycom_dum_with")
abline(0,0, col="red", lty=3, lwd=2)
abline(h=model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt$coefficients["psycom_dum_with"], col=3, lwd=2, lty= 3)
legend("bottomleft", legend= c("Reference line for the null effect", "Average hazard over time", "Time-varying hazard"), lty=c(3,2,1), col= c("red",3,1), lwd=2, bty = "n", bg = "transparent")
#autoplot(km_home) # just to see the km curves
km_psycom_dum_study_restr <- survfit(Surv(lag_time,time,event)~ psycom_dum_study+
cluster(id), data = data_mine_miss_restr_proc2)
plot(km_psycom_dum_study_restr, fun = "cloglog", xlab = "Time (in days) using log",col=c("black", "red"), ylab = "log-log survival", main = "log-log curves by psycom_dum_study (restricted)")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1, bty = "n", bg = "transparent")
km_psycom_dum_with_restr <- survfit(Surv(lag_time,time,event)~ psycom_dum_with+
cluster(id), data = data_mine_miss_restr_proc2)
#autoplot(km_home) # just to see the km curves
plot(km_psycom_dum_with_restr, fun = "cloglog", xlab = "Time (in days) using log",col=c("black", "red"), ylab = "log-log survival", main = "log-log curves by km_psycom_dum_with (restricted)")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1, bty = "n", bg = "transparent")
Code
recorded_plot <- recordPlot()
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
png(paste0(folder_path,"zph/proportionallity_psycom_restr.png"), height=13, width=10, res=500, units="in")
recorded_plot
dev.off()png
2 Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("2024-05-01: restricted")
res240422_4_restr <- resid(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt, "scaledsch")
time240422_4_restr <- as.numeric(dimnames(res240422_3_restr)[[1]])
z_240422_4_1_restr <- loess(res240422_4_restr[,"lag_tr_outcome"] ~ time240422_4_restr, span=0.50) # residuals
# x11()
par(mfrow= c(3,1))
plot(time240422_4_restr, fitted(z_240422_4_1_restr))
lines(supsmu(time240422_4_restr, res240422_4_restr[,"lag_tr_outcome"]),lty=2)
title("Sch. res., lag_tr_outcome")
plot(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt)[which(attr(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt$coefficients, "names") == "lag_tr_outcome")], lwd=2, main= "Residuals of lag_tr_outcome")
abline(0,0, col="red", lty=3, lwd=2)
abline(h=model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt$coefficients["lag_tr_outcome"], col=3, lwd=2, lty= 3)
legend("bottomleft", legend= c("Reference line for the null effect", "Average hazard over time", "Time-varying hazard"), lty=c(3,2,1), col= c("red",3,1), lwd=2, bty = "n", bg = "transparent")
#autoplot(km_home) # just to see the km curves
km_lag_tr_outcome_restr <- survfit(Surv(lag_time,time,event)~ lag_tr_outcome+
cluster(id), data = data_mine_miss_restr_proc2)
plot(km_lag_tr_outcome_restr, fun = "cloglog", xlab = "Time (in days) using log",col=c("black", "red"), ylab = "log-log survival", main = "log-log curves by km_lag_tr_outcome (restricted)")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1, bty = "n", bg = "transparent")
Code
recorded_plot <- recordPlot()
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
png(paste0(folder_path,"zph/proportionallity_lag_tr_outcome_restr.png"), height=13, width=10, res=500, units="in")
recorded_plot
dev.off()png
2 Code
#load("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/20240502_avance_iiw.RData")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_lag_tr_outcome <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq(.1,1.5,.025)) {
tvcs <- c(20,80)*i
data_mine_miss_restr_proc2$lag_tr_outcome_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$lag_tr_outcome,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
data_mine_miss_restr_proc2$tr_outcome_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$tr_outcome,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)
invisible("")
print(paste0("iteration: ",i))
aic_iter <- round(extractAIC(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)[[2]],1)
print( paste0("Time points: ",tvcs[1],", ", tvcs[2]))
print( paste0("AIC: ", aic_iter))
cox_zph_iter <- format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)$table), 1])
print(paste0("Cox ZPH: ", cox_zph_iter))
results_lag_tr_outcome <- rbind.data.frame(results_lag_tr_outcome,
cbind.data.frame(
spec = paste0("iteration: ",i, "; tvcs: ", tvcs[1],", ", tvcs[2]),
aic = aic_iter,
cox_zph = cox_zph_iter,
success = T
))
}[1] "iteration: 0.1"
[1] "Time points: 2, 8"
[1] "AIC: 241646.4"
[1] "Cox ZPH: 6034.712"
[1] "iteration: 0.125"
[1] "Time points: 2.5, 10"
[1] "AIC: 241490.1"
[1] "Cox ZPH: 6318.221"
[1] "iteration: 0.15"
[1] "Time points: 3, 12"
[1] "AIC: 241364.7"
[1] "Cox ZPH: 6459.451"
[1] "iteration: 0.175"
[1] "Time points: 3.5, 14"
[1] "AIC: 241275"
[1] "Cox ZPH: 6445.888"
[1] "iteration: 0.2"
[1] "Time points: 4, 16"
[1] "AIC: 241227.8"
[1] "Cox ZPH: 6262.837"
[1] "iteration: 0.225"
[1] "Time points: 4.5, 18"
[1] "AIC: 241218.5"
[1] "Cox ZPH: 5937.099"
[1] "iteration: 0.25"
[1] "Time points: 5, 20"
[1] "AIC: 241244"
[1] "Cox ZPH: 5480.285"
[1] "iteration: 0.275"
[1] "Time points: 5.5, 22"
[1] "AIC: 241314.3"
[1] "Cox ZPH: 4871.568"
[1] "iteration: 0.3"
[1] "Time points: 6, 24"
[1] "AIC: 241431.2"
[1] "Cox ZPH: 4108.697"
[1] "iteration: 0.325"
[1] "Time points: 6.5, 26"
[1] "AIC: 241579.4"
[1] "Cox ZPH: 3270.654"
[1] "iteration: 0.35"
[1] "Time points: 7, 28"
[1] "AIC: 241763.2"
[1] "Cox ZPH: 2385.785"
[1] "iteration: 0.375"
[1] "Time points: 7.5, 30"
[1] "AIC: 241978.3"
[1] "Cox ZPH: 1540.282"
[1] "iteration: 0.4"
[1] "Time points: 8, 32"
[1] "AIC: 242214.4"
[1] "Cox ZPH: 856.6484"
[1] "iteration: 0.425"
[1] "Time points: 8.5, 34"
[1] "AIC: 242469.5"
[1] "Cox ZPH: 423.423"
[1] "iteration: 0.45"
[1] "Time points: 9, 36"
[1] "AIC: 242741.2"
[1] "Cox ZPH: 321.6527"
[1] "iteration: 0.475"
[1] "Time points: 9.5, 38"
[1] "AIC: 243020.3"
[1] "Cox ZPH: 600.673"
[1] "iteration: 0.5"
[1] "Time points: 10, 40"
[1] "AIC: 243294.4"
[1] "Cox ZPH: 1241.962"
[1] "iteration: 0.525"
[1] "Time points: 10.5, 42"
[1] "AIC: 243564.4"
[1] "Cox ZPH: 2200.14"
[1] "iteration: 0.55"
[1] "Time points: 11, 44"
[1] "AIC: 243826.5"
[1] "Cox ZPH: 3401.45"
[1] "iteration: 0.575"
[1] "Time points: 11.5, 46"
[1] "AIC: 244069.7"
[1] "Cox ZPH: 4719.541"
[1] "iteration: 0.6"
[1] "Time points: 12, 48"
[1] "AIC: 244289.8"
[1] "Cox ZPH: 6050.557"
[1] "iteration: 0.625"
[1] "Time points: 12.5, 50"
[1] "AIC: 244485"
[1] "Cox ZPH: 7321.242"
[1] "iteration: 0.65"
[1] "Time points: 13, 52"
[1] "AIC: 244649.6"
[1] "Cox ZPH: 8448.845"
[1] "iteration: 0.675"
[1] "Time points: 13.5, 54"
[1] "AIC: 244779.6"
[1] "Cox ZPH: 9389.054"
[1] "iteration: 0.7"
[1] "Time points: 14, 56"
[1] "AIC: 244870"
[1] "Cox ZPH: 10113.72"
[1] "iteration: 0.725"
[1] "Time points: 14.5, 58"
[1] "AIC: 244916.8"
[1] "Cox ZPH: 10634.71"
[1] "iteration: 0.75"
[1] "Time points: 15, 60"
[1] "AIC: 244917.4"
[1] "Cox ZPH: 10977.17"
[1] "iteration: 0.775"
[1] "Time points: 15.5, 62"
[1] "AIC: 244869.9"
[1] "Cox ZPH: 11163.84"
[1] "iteration: 0.8"
[1] "Time points: 16, 64"
[1] "AIC: 244772.8"
[1] "Cox ZPH: 11235.97"
[1] "iteration: 0.825"
[1] "Time points: 16.5, 66"
[1] "AIC: 244626.8"
[1] "Cox ZPH: 11225.28"
[1] "iteration: 0.85"
[1] "Time points: 17, 68"
[1] "AIC: 244433.5"
[1] "Cox ZPH: 11152.15"
[1] "iteration: 0.875"
[1] "Time points: 17.5, 70"
[1] "AIC: 244193.4"
[1] "Cox ZPH: 11031.52"
[1] "iteration: 0.9"
[1] "Time points: 18, 72"
[1] "AIC: 243905.5"
[1] "Cox ZPH: 10877.46"
[1] "iteration: 0.925"
[1] "Time points: 18.5, 74"
[1] "AIC: 243572.4"
[1] "Cox ZPH: 10696.07"
[1] "iteration: 0.95"
[1] "Time points: 19, 76"
[1] "AIC: 243196.2"
[1] "Cox ZPH: 10495.89"
[1] "iteration: 0.975"
[1] "Time points: 19.5, 78"
[1] "AIC: 242775.1"
[1] "Cox ZPH: 10286.2"
[1] "iteration: 1"
[1] "Time points: 20, 80"
[1] "AIC: 242307"
[1] "Cox ZPH: 10071.06"
[1] "iteration: 1.025"
[1] "Time points: 20.5, 82"
[1] "AIC: 241797.4"
[1] "Cox ZPH: 9853.469"
[1] "iteration: 1.05"
[1] "Time points: 21, 84"
[1] "AIC: 241247"
[1] "Cox ZPH: 9636.775"
[1] "iteration: 1.075"
[1] "Time points: 21.5, 86"
[1] "AIC: 240656.4"
[1] "Cox ZPH: 9416.648"
[1] "iteration: 1.1"
[1] "Time points: 22, 88"
[1] "AIC: 240017.6"
[1] "Cox ZPH: 9186.961"
[1] "iteration: 1.125"
[1] "Time points: 22.5, 90"
[1] "AIC: 239332.2"
[1] "Cox ZPH: 8946.098"
[1] "iteration: 1.15"
[1] "Time points: 23, 92"
[1] "AIC: 238596"
[1] "Cox ZPH: 8687.54"
[1] "iteration: 1.175"
[1] "Time points: 23.5, 94"
[1] "AIC: 237808.6"
[1] "Cox ZPH: 8407.242"
[1] "iteration: 1.2"
[1] "Time points: 24, 96"
[1] "AIC: 236965.9"
[1] "Cox ZPH: 8104.911"
[1] "iteration: 1.225"
[1] "Time points: 24.5, 98"
[1] "AIC: 236068.8"
[1] "Cox ZPH: 7780.982"
[1] "iteration: 1.25"
[1] "Time points: 25, 100"
[1] "AIC: 235123.9"
[1] "Cox ZPH: 7435.731"
[1] "iteration: 1.275"
[1] "Time points: 25.5, 102"
[1] "AIC: 234128.5"
[1] "Cox ZPH: 7071.559"
[1] "iteration: 1.3"
[1] "Time points: 26, 104"
[1] "AIC: 233087.8"
[1] "Cox ZPH: 6690.547"
[1] "iteration: 1.325"
[1] "Time points: 26.5, 106"
[1] "AIC: 232023.4"
[1] "Cox ZPH: 6300.418"
[1] "iteration: 1.35"
[1] "Time points: 27, 108"
[1] "AIC: 230934.8"
[1] "Cox ZPH: 5901.547"
[1] "iteration: 1.375"
[1] "Time points: 27.5, 110"
[1] "AIC: 229821"
[1] "Cox ZPH: 5497.145"
[1] "iteration: 1.4"
[1] "Time points: 28, 112"
[1] "AIC: 228696.5"
[1] "Cox ZPH: 5091.223"
[1] "iteration: 1.425"
[1] "Time points: 28.5, 114"
[1] "AIC: 227581"
[1] "Cox ZPH: 4677.408"
[1] "iteration: 1.45"
[1] "Time points: 29, 116"
[1] "AIC: 226445.4"
[1] "Cox ZPH: 4221.169"
[1] "iteration: 1.475"
[1] "Time points: 29.5, 118"
[1] "AIC: 225226.2"
[1] "Cox ZPH: 3672.86"
[1] "iteration: 1.5"
[1] "Time points: 30, 120"
[1] "AIC: 223853.3"
[1] "Cox ZPH: 2998.596"Code
results_lag_tr_outcome[which(as.numeric(results_lag_tr_outcome$cox_zph)==
min(as.numeric(results_lag_tr_outcome$cox_zph))),] spec aic cox_zph success
15 iteration: 0.45; tvcs: 9, 36 242741.2 321.6527 TRUECode
# spec aic cox_zph success
# 15 iteration: 0.45; tvcs: 9, 36 242729.5 321.3004 TRUE
message(paste0("The previously modified model: ",
round(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_alt)$table), 1],1),
" vs. the minimum model: ", results_lag_tr_outcome[which(as.numeric(results_lag_tr_outcome$cox_zph)== min(as.numeric(results_lag_tr_outcome$cox_zph))),"cox_zph"]
))Code
#The previously modified model: 361.9 vs. the minimum model: 321.3004
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
require(splines)
specs_lag_tr_out <- paste0("lag_tr_outcome *", c("time", "log(time+0.001)", "poly(ifelse(is.na(time), .001,time), 3)", "poly(ifelse(is.na(time), .001,time), 4)", "poly(ifelse(is.na(time), .001,time), 5)", "poly(ifelse(is.na(time), .001,time), 6)", "pspline(log(time+0.001), df=3)", "pspline(log(time+0.001), df=4)", "pspline(log(time+0.001), df=5)", "pspline(log(time+0.001), df=6)", "bs(time, 3)", "bs(time, 4)", "bs(time, 5)", "bs(time, 6)", "ns(time, 3)", "ns(time, 4)", "ns(time, 5)", "ns(time, 6)", "rcs(time, 3)", "rcs(time, 4)", "rcs(time, 5)", "rcs(time, 6)"))
results_lag_tr_out <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq_along(specs_lag_tr_out)) {
l <- specs_lag_tr_out[i]
result <- tryCatch({
model <- cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
I(round(",l,",2)) +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
dplyr::mutate(log_lag_dias_treat_imp_sin_na_rec = bs(log_lag_dias_treat_imp_sin_na, 3)) %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
aic_val <- extractAIC(model)[[2]]
cox_zph_val <- cox.zph(model)$table[nrow(cox.zph(model)$table), 1]
success <- TRUE
}, error = function(e) {
aic_val <- NA
cox_zph_val <- NA
success <- FALSE
warning(paste0("Error fitting specification ", i, " (", l, "): ", e))
})
results_lag_tr_out <- rbind(results_lag_tr_out, data.frame(spec = l, aic = aic_val, cox_zph = cox_zph_val, success = success))
print(cat(paste("Iteration:", i, "\n"))) # Add iteration number
print(cat(paste("Spec: ", l, "\n")))
print(cat(paste("AIC: ", results_lag_tr_out$aic[i], "\n")))
print(cat(paste("Cox zph test statistic: ", results_lag_tr_out$cox_zph[i], "\n")))
}Iteration: 1
NULL
Spec: lag_tr_outcome *time
NULL
AIC: 242375.661429718
NULL
Cox zph test statistic: 4294.36756065618
NULL
Iteration: 2
NULL
Spec: lag_tr_outcome *log(time+0.001)
NULL
AIC: 244573.478733201
NULL
Cox zph test statistic: 1250.69604172358
NULL
Iteration: 3
NULL
Spec: lag_tr_outcome *poly(ifelse(is.na(time), .001,time), 3)
NULL
AIC: 241209.963903167
NULL
Cox zph test statistic: 12162.2885407214
NULL
Iteration: 4
NULL
Spec: lag_tr_outcome *poly(ifelse(is.na(time), .001,time), 4)
NULL
AIC: 240784.062529073
NULL
Cox zph test statistic: 12595.5788006453
NULL
Iteration: 5
NULL
Spec: lag_tr_outcome *poly(ifelse(is.na(time), .001,time), 5)
NULL
AIC: 240655.695209592
NULL
Cox zph test statistic: 12721.8186063262
NULL
Iteration: 6
NULL
Spec: lag_tr_outcome *poly(ifelse(is.na(time), .001,time), 6)
NULL
AIC: 240210.222768023
NULL
Cox zph test statistic: 12447.9992558718
NULL
Iteration: 7
NULL
Spec: lag_tr_outcome *pspline(log(time+0.001), df=3)
NULL
AIC: 240210.222768023
NULL
Cox zph test statistic: 12447.9992558718
NULL
Iteration: 8
NULL
Spec: lag_tr_outcome *pspline(log(time+0.001), df=4)
NULL
AIC: 240210.222768023
NULL
Cox zph test statistic: 12447.9992558718
NULL
Iteration: 9
NULL
Spec: lag_tr_outcome *pspline(log(time+0.001), df=5)
NULL
AIC: 240210.222768023
NULL
Cox zph test statistic: 12447.9992558718
NULL
Iteration: 10
NULL
Spec: lag_tr_outcome *pspline(log(time+0.001), df=6)
NULL
AIC: 240210.222768023
NULL
Cox zph test statistic: 12447.9992558718
NULL
Iteration: 11
NULL
Spec: lag_tr_outcome *bs(time, 3)
NULL
AIC: 238139.168565678
NULL
Cox zph test statistic: 20194.0931692
NULL
Iteration: 12
NULL
Spec: lag_tr_outcome *bs(time, 4)
NULL
AIC: 221388.904711619
NULL
Cox zph test statistic: 92983.6270487867
NULL
Iteration: 13
NULL
Spec: lag_tr_outcome *bs(time, 5)
NULL
AIC: 213169.70279823
NULL
Cox zph test statistic: 77263.2011050867
NULL
Iteration: 14
NULL
Spec: lag_tr_outcome *bs(time, 6)
NULL
AIC: 211342.28883079
NULL
Cox zph test statistic: 103010.888687678
NULL
Iteration: 15
NULL
Spec: lag_tr_outcome *ns(time, 3)
NULL
AIC: 230455.219132658
NULL
Cox zph test statistic: 50605.4601854322
NULL
Iteration: 16
NULL
Spec: lag_tr_outcome *ns(time, 4)
NULL
AIC: 219389.439286971
NULL
Cox zph test statistic: 78509.834057776
NULL
Iteration: 17
NULL
Spec: lag_tr_outcome *ns(time, 5)
NULL
AIC: 215468.139506862
NULL
Cox zph test statistic: 83287.572463163
NULL
Iteration: 18
NULL
Spec: lag_tr_outcome *ns(time, 6)
NULL
AIC: 213583.907149193
NULL
Cox zph test statistic: 101262.397922225
NULL
Iteration: 19
NULL
Spec: lag_tr_outcome *rcs(time, 3)
NULL
AIC: 239536.406693165
NULL
Cox zph test statistic: 17245.5074031239
NULL
Iteration: 20
NULL
Spec: lag_tr_outcome *rcs(time, 4)
NULL
AIC: 228684.622158062
NULL
Cox zph test statistic: 65410.1688298938
NULL
Iteration: 21
NULL
Spec: lag_tr_outcome *rcs(time, 5)
NULL
AIC: 226434.369582259
NULL
Cox zph test statistic: 23238.4847315333
NULL
X matrix deemed to be singular; variable time'''
Iteration: 22
NULL
Spec: lag_tr_outcome *rcs(time, 6)
NULL
AIC: 226434.369582259
NULL
Cox zph test statistic: 23238.4847315333
NULLCode
results_lag_tr_out[which(as.numeric(results_lag_tr_out$cox_zph)==
min(as.numeric(results_lag_tr_out$cox_zph))),] spec aic cox_zph success
2 lag_tr_outcome *log(time+0.001) 244573.5 1250.696 TRUECode
# spec aic cox_zph success
# 2 lag_tr_outcome *log(time+0.001) 244561.9 1248.051 TRUE
invisible("It worsen. We do nothing")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_lag_tr_outcome_alt <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq(1,5,.1)) {
change_points <- c(5*i, 10*i, 15*i) # Example change points (months)
coefficients <- c(0.5, 1.5, 0.5, 1) # Example coefficients
# Add the time-varying coefficient to your model
data_mine_miss_restr_proc2$lag_tr_outcome_rec2 <- with(data_mine_miss_restr_proc2,
lag_tr_outcome*time_varying_coeff(lag_time, change_points, coefficients))
data_mine_miss_restr_proc2$tr_outcome_rec2 <- with(data_mine_miss_restr_proc2,
tr_outcome*time_varying_coeff(time, change_points, coefficients))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_alt<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec2 +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_alt)
invisible("")
print(paste0("iteration: ",i))
aic_iter <- round(extractAIC(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_alt)[[2]],1)
print( paste0("Time points: ",change_points[1],", ", change_points[2], ", ",change_points[3]))
print( paste0("AIC: ", aic_iter))
cox_zph_iter <- format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_alt)$table), 1])
print(paste0("Cox ZPH: ", cox_zph_iter))
results_lag_tr_outcome_alt <- rbind.data.frame(results_lag_tr_outcome_alt,
cbind.data.frame(
spec = paste0("iteration: ",i, "; change points: ", change_points[1],", ", change_points[2], ", ",change_points[3]),
aic = aic_iter,
cox_zph = cox_zph_iter,
success = T
))
}[1] "iteration: 1"
[1] "Time points: 5, 10, 15"
[1] "AIC: 244899.5"
[1] "Cox ZPH: 359.0971"
[1] "iteration: 1.1"
[1] "Time points: 5.5, 11, 16.5"
[1] "AIC: 244900.4"
[1] "Cox ZPH: 361.2743"
[1] "iteration: 1.2"
[1] "Time points: 6, 12, 18"
[1] "AIC: 244903.3"
[1] "Cox ZPH: 358.6406"
[1] "iteration: 1.3"
[1] "Time points: 6.5, 13, 19.5"
[1] "AIC: 244906.3"
[1] "Cox ZPH: 359.1773"
[1] "iteration: 1.4"
[1] "Time points: 7, 14, 21"
[1] "AIC: 244906.5"
[1] "Cox ZPH: 360.6033"
[1] "iteration: 1.5"
[1] "Time points: 7.5, 15, 22.5"
[1] "AIC: 244906.8"
[1] "Cox ZPH: 362.1983"
[1] "iteration: 1.6"
[1] "Time points: 8, 16, 24"
[1] "AIC: 244902.5"
[1] "Cox ZPH: 361.868"
[1] "iteration: 1.7"
[1] "Time points: 8.5, 17, 25.5"
[1] "AIC: 244901.8"
[1] "Cox ZPH: 364.0377"
[1] "iteration: 1.8"
[1] "Time points: 9, 18, 27"
[1] "AIC: 244905.9"
[1] "Cox ZPH: 364.3575"
[1] "iteration: 1.9"
[1] "Time points: 9.5, 19, 28.5"
[1] "AIC: 244905.3"
[1] "Cox ZPH: 365.32"
[1] "iteration: 2"
[1] "Time points: 10, 20, 30"
[1] "AIC: 244901.4"
[1] "Cox ZPH: 368.9659"
[1] "iteration: 2.1"
[1] "Time points: 10.5, 21, 31.5"
[1] "AIC: 244901.9"
[1] "Cox ZPH: 369.6159"
[1] "iteration: 2.2"
[1] "Time points: 11, 22, 33"
[1] "AIC: 244903"
[1] "Cox ZPH: 365.9987"
[1] "iteration: 2.3"
[1] "Time points: 11.5, 23, 34.5"
[1] "AIC: 244895.1"
[1] "Cox ZPH: 361.7019"
[1] "iteration: 2.4"
[1] "Time points: 12, 24, 36"
[1] "AIC: 244896.9"
[1] "Cox ZPH: 356.4047"
[1] "iteration: 2.5"
[1] "Time points: 12.5, 25, 37.5"
[1] "AIC: 244893.9"
[1] "Cox ZPH: 357.1797"
[1] "iteration: 2.6"
[1] "Time points: 13, 26, 39"
[1] "AIC: 244891.6"
[1] "Cox ZPH: 356.8042"
[1] "iteration: 2.7"
[1] "Time points: 13.5, 27, 40.5"
[1] "AIC: 244885.7"
[1] "Cox ZPH: 354.4202"
[1] "iteration: 2.8"
[1] "Time points: 14, 28, 42"
[1] "AIC: 244880.4"
[1] "Cox ZPH: 354.3132"
[1] "iteration: 2.9"
[1] "Time points: 14.5, 29, 43.5"
[1] "AIC: 244879.2"
[1] "Cox ZPH: 352.0868"
[1] "iteration: 3"
[1] "Time points: 15, 30, 45"
[1] "AIC: 244879.9"
[1] "Cox ZPH: 349.6504"
[1] "iteration: 3.1"
[1] "Time points: 15.5, 31, 46.5"
[1] "AIC: 244883.7"
[1] "Cox ZPH: 349.1183"
[1] "iteration: 3.2"
[1] "Time points: 16, 32, 48"
[1] "AIC: 244880.6"
[1] "Cox ZPH: 348.1966"
[1] "iteration: 3.3"
[1] "Time points: 16.5, 33, 49.5"
[1] "AIC: 244877.7"
[1] "Cox ZPH: 348.1207"
[1] "iteration: 3.4"
[1] "Time points: 17, 34, 51"
[1] "AIC: 244881.7"
[1] "Cox ZPH: 348.0226"
[1] "iteration: 3.5"
[1] "Time points: 17.5, 35, 52.5"
[1] "AIC: 244881.8"
[1] "Cox ZPH: 348.0872"
[1] "iteration: 3.6"
[1] "Time points: 18, 36, 54"
[1] "AIC: 244883"
[1] "Cox ZPH: 349.6944"
[1] "iteration: 3.7"
[1] "Time points: 18.5, 37, 55.5"
[1] "AIC: 244883.3"
[1] "Cox ZPH: 350.0009"
[1] "iteration: 3.8"
[1] "Time points: 19, 38, 57"
[1] "AIC: 244880.6"
[1] "Cox ZPH: 350.4467"
[1] "iteration: 3.9"
[1] "Time points: 19.5, 39, 58.5"
[1] "AIC: 244877.4"
[1] "Cox ZPH: 350.8478"
[1] "iteration: 4"
[1] "Time points: 20, 40, 60"
[1] "AIC: 244878.7"
[1] "Cox ZPH: 351.8303"
[1] "iteration: 4.1"
[1] "Time points: 20.5, 41, 61.5"
[1] "AIC: 244880"
[1] "Cox ZPH: 354.9833"
[1] "iteration: 4.2"
[1] "Time points: 21, 42, 63"
[1] "AIC: 244877"
[1] "Cox ZPH: 354.6689"
[1] "iteration: 4.3"
[1] "Time points: 21.5, 43, 64.5"
[1] "AIC: 244873.9"
[1] "Cox ZPH: 355.1347"
[1] "iteration: 4.4"
[1] "Time points: 22, 44, 66"
[1] "AIC: 244871.1"
[1] "Cox ZPH: 356.1269"
[1] "iteration: 4.5"
[1] "Time points: 22.5, 45, 67.5"
[1] "AIC: 244874.5"
[1] "Cox ZPH: 357.1391"
[1] "iteration: 4.6"
[1] "Time points: 23, 46, 69"
[1] "AIC: 244875.8"
[1] "Cox ZPH: 358.4438"
[1] "iteration: 4.7"
[1] "Time points: 23.5, 47, 70.5"
[1] "AIC: 244874.8"
[1] "Cox ZPH: 359.9124"
[1] "iteration: 4.8"
[1] "Time points: 24, 48, 72"
[1] "AIC: 244865.4"
[1] "Cox ZPH: 358.8837"
[1] "iteration: 4.9"
[1] "Time points: 24.5, 49, 73.5"
[1] "AIC: 244868.3"
[1] "Cox ZPH: 360.8408"
[1] "iteration: 5"
[1] "Time points: 25, 50, 75"
[1] "AIC: 244868.9"
[1] "Cox ZPH: 361.8831"Code
results_lag_tr_outcome_alt[which(as.numeric(results_lag_tr_outcome_alt$cox_zph)==
min(as.numeric(results_lag_tr_outcome_alt$cox_zph))),] spec aic cox_zph success
25 iteration: 3.4; change points: 17, 34, 51 244881.7 348.0226 TRUECode
# spec aic cox_zph success
# 25 iteration: 3.4; change points: 17, 34, 51 244869.2 347.3415 TRUE
message(paste0("The minimum of the first modified model: ",
results_lag_tr_outcome[which(as.numeric(results_lag_tr_outcome$cox_zph)== min(as.numeric(results_lag_tr_outcome$cox_zph))),"cox_zph"],
" vs. the minimum model: ", results_lag_tr_outcome_alt[which(as.numeric(results_lag_tr_outcome_alt$cox_zph)== min(as.numeric(results_lag_tr_outcome_alt$cox_zph))),"cox_zph"]
))Code
#The minimum of the first modified model: 321.3004 vs. the minimum model: 347.3415
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_lag_tr_outcome_alt[which(as.numeric(results_lag_tr_outcome$cox_zph)==
min(as.numeric(results_lag_tr_outcome$cox_zph))),] spec aic cox_zph success
15 iteration: 2.4; change points: 12, 24, 36 244896.9 356.4047 TRUECode
# spec aic cox_zph success
# 15 iteration: 0.45; tvcs: 9, 36 242729.5 321.3004 TRUE
tvcs<- c(9, 36)
data_mine_miss_restr_proc2$lag_tr_outcome_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$lag_tr_outcome,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
data_mine_miss_restr_proc2$tr_outcome_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$tr_outcome,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out<-
cph(Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
round(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)$table,4) chisq df p
lag_tr_outcome_rec 66.5194 1 0.0000
log_lag_dias_treat_imp_sin_na 10.3514 1 0.0013
lag_less_90d_tr1_rec 84.0336 1 0.0000
lag_comp_bpsc_y3_severe_rec 17.2563 1 0.0000
lag_policonsumo2 3.2167 1 0.0729
edad_al_ing_1 0.1446 1 0.7038
ano_nac_corr 5.1810 1 0.0228
susinidum_oh 20.1319 1 0.0000
susinidum_coc 9.7922 1 0.0018
susinidum_pbc 2.0721 1 0.1500
susinidum_mar 6.7050 1 0.0096
psycom_dum_with 11.7329 1 0.0006
psycom_dum_study 1.6100 1 0.2045
freq_cons_dum_5day 0.0801 1 0.7772
cond_oc_dum_2inact 4.4364 1 0.0352
cond_oc_dum_3unemp 1.7433 1 0.1867
susprindum_oh 4.4295 1 0.0353
susprindum_coc 1.6332 1 0.2013
susprindum_pbc 0.0553 1 0.8141
susprindum_mar 5.2895 1 0.0215
GLOBAL 321.6527 20 0.0000Code
# chisq df p
# lag_tr_outcome_rec 66.6269 1 0.0000 ####
# log_lag_dias_treat_imp_sin_na 10.1949 1 0.0014 #
# lag_less_90d_tr1_rec 84.0361 1 0.0000 ####
# lag_comp_bpsc_y3_severe_rec 17.2093 1 0.0000 ####
# lag_policonsumo2 3.1849 1 0.0743
# edad_al_ing_1 0.1368 1 0.7115
# ano_nac_corr 5.2180 1 0.0224 #
# susinidum_oh 19.9277 1 0.0000 ####
# susinidum_coc 9.7729 1 0.0018 ##
# susinidum_pbc 1.9935 1 0.1580
# susinidum_mar 6.6712 1 0.0098 #
# psycom_dum_with 11.7874 1 0.0006 ##
# psycom_dum_study 1.6215 1 0.2029
# freq_cons_dum_5day 0.0521 1 0.8194
# cond_oc_dum_2inact 4.4607 1 0.0347
# cond_oc_dum_3unemp 1.6238 1 0.2026
# susprindum_oh 4.3416 1 0.0372
# susprindum_coc 1.6713 1 0.1961
# susprindum_pbc 0.0723 1 0.7879
# susprindum_mar 5.3392 1 0.0209
# GLOBAL 321.3004 20 0.0000
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Analyse the violation of proportional hazards")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
try(dev.off())null device
1 Code
res240422_5_restr <- resid(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out, "scaledsch")
time240422_5_restr <- as.numeric(dimnames(res240422_5_restr)[[1]])
z_240422_5_1_restr <- loess(res240422_5_restr[,"susinidum_oh"] ~ time240422_5_restr, span=0.50) # residuals
par(mfrow= c(3,1))
plot(time240422_5_restr, fitted(z_240422_5_1_restr))
lines(supsmu(time240422_5_restr, res240422_5_restr[,"susinidum_oh"]),lty=2)
title("Sch. res., susinidum_oh")
plot(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)[which(attr(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out$coefficients, "names") == "susinidum_oh")], lwd=2, main= "Residuals of susinidum_oh")
abline(0,0, col="red", lty=3, lwd=2)
abline(h=model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out$coefficients["susinidum_oh"], col=3, lwd=2, lty= 3)
legend("bottomleft", legend= c("Reference line for the null effect", "Average hazard over time", "Time-varying hazard"), lty=c(3,2,1), col= c("red",3,1), lwd=2, bty = "n", bg = "transparent")
km_susinidum_oh <- survfit(Surv(lag_time,time,event)~ susinidum_oh+
cluster(id), data = data_mine_miss_restr_proc2)
plot(km_susinidum_oh, fun = "cloglog", xlab = "Time (in days) using log",col=c("black", "red"), ylab = "log-log survival", main = "log-log curves by susinidum_oh")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1, bty = "n", bg = "transparent")
recorded_plot <- recordPlot()
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
png(paste0(folder_path,"zph/proportionallity_susinidum_oh_restr.png"), height=13, width=10, res=500, units="in")
recorded_plot
dev.off()pdf
2 Code
invisible("Now with Initial substance: Alcohol")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_susinidum_oh <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq(.1,1.5,.025)) {
tvcs <- c(20,80)*i
data_mine_miss_restr_proc2$susinidum_oh_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$susinidum_oh,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh_rec +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh)
invisible("")
print(paste0("iteration: ",i))
aic_iter <- round(extractAIC(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh)[[2]],1)
print( paste0("Time points: ",tvcs[1],", ", tvcs[2]))
print( paste0("AIC: ", aic_iter))
cox_zph_iter <- format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh)$table), 1])
print(paste0("Cox ZPH: ", cox_zph_iter))
results_susinidum_oh <- rbind.data.frame(results_susinidum_oh,
cbind.data.frame(
spec = paste0("iteration: ",i, "; tvcs: ", tvcs[1],", ", tvcs[2]),
aic = aic_iter,
cox_zph = cox_zph_iter,
success = T
))
}[1] "iteration: 0.1"
[1] "Time points: 2, 8"
[1] "AIC: 239429.8"
[1] "Cox ZPH: 5741.105"
[1] "iteration: 0.125"
[1] "Time points: 2.5, 10"
[1] "AIC: 239392.7"
[1] "Cox ZPH: 5911.203"
[1] "iteration: 0.15"
[1] "Time points: 3, 12"
[1] "AIC: 239385.1"
[1] "Cox ZPH: 5995.153"
[1] "iteration: 0.175"
[1] "Time points: 3.5, 14"
[1] "AIC: 239405.8"
[1] "Cox ZPH: 6004.474"
[1] "iteration: 0.2"
[1] "Time points: 4, 16"
[1] "AIC: 239458"
[1] "Cox ZPH: 5931.333"
[1] "iteration: 0.225"
[1] "Time points: 4.5, 18"
[1] "AIC: 239533.6"
[1] "Cox ZPH: 5798.068"
[1] "iteration: 0.25"
[1] "Time points: 5, 20"
[1] "AIC: 239629.6"
[1] "Cox ZPH: 5605.943"
[1] "iteration: 0.275"
[1] "Time points: 5.5, 22"
[1] "AIC: 239746.7"
[1] "Cox ZPH: 5345.649"
[1] "iteration: 0.3"
[1] "Time points: 6, 24"
[1] "AIC: 239888"
[1] "Cox ZPH: 4997.661"
[1] "iteration: 0.325"
[1] "Time points: 6.5, 26"
[1] "AIC: 240042.6"
[1] "Cox ZPH: 4586.522"
[1] "iteration: 0.35"
[1] "Time points: 7, 28"
[1] "AIC: 240211.5"
[1] "Cox ZPH: 4104.321"
[1] "iteration: 0.375"
[1] "Time points: 7.5, 30"
[1] "AIC: 240387.4"
[1] "Cox ZPH: 3584.708"
[1] "iteration: 0.4"
[1] "Time points: 8, 32"
[1] "AIC: 240565.7"
[1] "Cox ZPH: 3057.196"
[1] "iteration: 0.425"
[1] "Time points: 8.5, 34"
[1] "AIC: 240747.1"
[1] "Cox ZPH: 2541.069"
[1] "iteration: 0.45"
[1] "Time points: 9, 36"
[1] "AIC: 240930.5"
[1] "Cox ZPH: 2078.364"
[1] "iteration: 0.475"
[1] "Time points: 9.5, 38"
[1] "AIC: 241113.3"
[1] "Cox ZPH: 1728.213"
[1] "iteration: 0.5"
[1] "Time points: 10, 40"
[1] "AIC: 241288.4"
[1] "Cox ZPH: 1549.536"
[1] "iteration: 0.525"
[1] "Time points: 10.5, 42"
[1] "AIC: 241459.6"
[1] "Cox ZPH: 1566.979"
[1] "iteration: 0.55"
[1] "Time points: 11, 44"
[1] "AIC: 241626.8"
[1] "Cox ZPH: 1794.567"
[1] "iteration: 0.575"
[1] "Time points: 11.5, 46"
[1] "AIC: 241787.2"
[1] "Cox ZPH: 2219.701"
[1] "iteration: 0.6"
[1] "Time points: 12, 48"
[1] "AIC: 241939.2"
[1] "Cox ZPH: 2804.371"
[1] "iteration: 0.625"
[1] "Time points: 12.5, 50"
[1] "AIC: 242083.7"
[1] "Cox ZPH: 3505.791"
[1] "iteration: 0.65"
[1] "Time points: 13, 52"
[1] "AIC: 242219.9"
[1] "Cox ZPH: 4264.682"
[1] "iteration: 0.675"
[1] "Time points: 13.5, 54"
[1] "AIC: 242345.1"
[1] "Cox ZPH: 5015.719"
[1] "iteration: 0.7"
[1] "Time points: 14, 56"
[1] "AIC: 242456.5"
[1] "Cox ZPH: 5696.982"
[1] "iteration: 0.725"
[1] "Time points: 14.5, 58"
[1] "AIC: 242551.4"
[1] "Cox ZPH: 6259.37"
[1] "iteration: 0.75"
[1] "Time points: 15, 60"
[1] "AIC: 242628.8"
[1] "Cox ZPH: 6683.401"
[1] "iteration: 0.775"
[1] "Time points: 15.5, 62"
[1] "AIC: 242687.4"
[1] "Cox ZPH: 6956.512"
[1] "iteration: 0.8"
[1] "Time points: 16, 64"
[1] "AIC: 242725.2"
[1] "Cox ZPH: 7085.722"
[1] "iteration: 0.825"
[1] "Time points: 16.5, 66"
[1] "AIC: 242740.8"
[1] "Cox ZPH: 7091.269"
[1] "iteration: 0.85"
[1] "Time points: 17, 68"
[1] "AIC: 242733.4"
[1] "Cox ZPH: 6995.256"
[1] "iteration: 0.875"
[1] "Time points: 17.5, 70"
[1] "AIC: 242701.7"
[1] "Cox ZPH: 6819.604"
[1] "iteration: 0.9"
[1] "Time points: 18, 72"
[1] "AIC: 242645.4"
[1] "Cox ZPH: 6586.407"
[1] "iteration: 0.925"
[1] "Time points: 18.5, 74"
[1] "AIC: 242565.4"
[1] "Cox ZPH: 6318.07"
[1] "iteration: 0.95"
[1] "Time points: 19, 76"
[1] "AIC: 242461.6"
[1] "Cox ZPH: 6026.228"
[1] "iteration: 0.975"
[1] "Time points: 19.5, 78"
[1] "AIC: 242333.9"
[1] "Cox ZPH: 5721.657"
[1] "iteration: 1"
[1] "Time points: 20, 80"
[1] "AIC: 242181.2"
[1] "Cox ZPH: 5408.085"
[1] "iteration: 1.025"
[1] "Time points: 20.5, 82"
[1] "AIC: 242004.9"
[1] "Cox ZPH: 5098.267"
[1] "iteration: 1.05"
[1] "Time points: 21, 84"
[1] "AIC: 241807"
[1] "Cox ZPH: 4801.207"
[1] "iteration: 1.075"
[1] "Time points: 21.5, 86"
[1] "AIC: 241592"
[1] "Cox ZPH: 4522.922"
[1] "iteration: 1.1"
[1] "Time points: 22, 88"
[1] "AIC: 241359.7"
[1] "Cox ZPH: 4260.851"
[1] "iteration: 1.125"
[1] "Time points: 22.5, 90"
[1] "AIC: 241110.3"
[1] "Cox ZPH: 4013.58"
[1] "iteration: 1.15"
[1] "Time points: 23, 92"
[1] "AIC: 240849.8"
[1] "Cox ZPH: 3785.966"
[1] "iteration: 1.175"
[1] "Time points: 23.5, 94"
[1] "AIC: 240579.4"
[1] "Cox ZPH: 3575.447"
[1] "iteration: 1.2"
[1] "Time points: 24, 96"
[1] "AIC: 240299.9"
[1] "Cox ZPH: 3386.906"
[1] "iteration: 1.225"
[1] "Time points: 24.5, 98"
[1] "AIC: 240018.9"
[1] "Cox ZPH: 3223.323"
[1] "iteration: 1.25"
[1] "Time points: 25, 100"
[1] "AIC: 239740.7"
[1] "Cox ZPH: 3085.226"
[1] "iteration: 1.275"
[1] "Time points: 25.5, 102"
[1] "AIC: 239466.2"
[1] "Cox ZPH: 2973.787"
[1] "iteration: 1.3"
[1] "Time points: 26, 104"
[1] "AIC: 239195"
[1] "Cox ZPH: 2887.063"
[1] "iteration: 1.325"
[1] "Time points: 26.5, 106"
[1] "AIC: 238931.4"
[1] "Cox ZPH: 2829.198"
[1] "iteration: 1.35"
[1] "Time points: 27, 108"
[1] "AIC: 238672.7"
[1] "Cox ZPH: 2804.632"
[1] "iteration: 1.375"
[1] "Time points: 27.5, 110"
[1] "AIC: 238420.6"
[1] "Cox ZPH: 2817.714"
[1] "iteration: 1.4"
[1] "Time points: 28, 112"
[1] "AIC: 238188.2"
[1] "Cox ZPH: 2873.441"
[1] "iteration: 1.425"
[1] "Time points: 28.5, 114"
[1] "AIC: 237975.4"
[1] "Cox ZPH: 2981.972"
[1] "iteration: 1.45"
[1] "Time points: 29, 116"
[1] "AIC: 237779.2"
[1] "Cox ZPH: 3159.954"
[1] "iteration: 1.475"
[1] "Time points: 29.5, 118"
[1] "AIC: 237606.6"
[1] "Cox ZPH: 3430.731"
[1] "iteration: 1.5"
[1] "Time points: 30, 120"
[1] "AIC: 237464"
[1] "Cox ZPH: 3868.406"Code
results_susinidum_oh[which(as.numeric(results_susinidum_oh$cox_zph)==
min(as.numeric(results_susinidum_oh$cox_zph))),] spec aic cox_zph success
17 iteration: 0.5; tvcs: 10, 40 241288.4 1549.536 TRUECode
# spec aic cox_zph success
# 15 iteration: 0.45; tvcs: 9, 36 242729.5 321.3004 TRUE
message(paste0("The previously modified model: ",
round(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)$table), 1],1),
" vs. the minimum model: ", results_susinidum_oh[which(as.numeric(results_susinidum_oh$cox_zph)== min(as.numeric(results_susinidum_oh$cox_zph))),"cox_zph"]
))Code
#The previously modified model: 321.3 vs. the minimum model: 1547.708
invisible("It worsen")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
require(splines)
specs_susinidum_oh <- paste0("susinidum_oh *", c("time", "log(time+0.001)", "poly(ifelse(is.na(time), .001,time), 3)", "poly(ifelse(is.na(time), .001,time), 4)", "poly(ifelse(is.na(time), .001,time), 5)", "poly(ifelse(is.na(time), .001,time), 6)", "pspline(log(time+0.001), df=3)", "pspline(log(time+0.001), df=4)", "pspline(log(time+0.001), df=5)", "pspline(log(time+0.001), df=6)", "bs(time, 3)", "bs(time, 4)", "bs(time, 5)", "bs(time, 6)", "ns(time, 3)", "ns(time, 4)", "ns(time, 5)", "ns(time, 6)", "rcs(time, 3)", "rcs(time, 4)", "rcs(time, 5)", "rcs(time, 6)"))
results_susini_oh <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq_along(specs_susinidum_oh)) {
l <- specs_susinidum_oh[i]
result <- tryCatch({
model <- cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
I(round(",l,",2)) +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
dplyr::mutate(log_lag_dias_treat_imp_sin_na_rec = bs(log_lag_dias_treat_imp_sin_na, 3)) %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
aic_val <- extractAIC(model)[[2]]
cox_zph_val <- cox.zph(model)$table[nrow(cox.zph(model)$table), 1]
success <- TRUE
}, error = function(e) {
aic_val <- NA
cox_zph_val <- NA
success <- FALSE
warning(paste0("Error fitting specification ", i, " (", l, "): ", e))
})
results_susini_oh <- rbind(results_susini_oh, data.frame(spec = l, aic = aic_val, cox_zph = cox_zph_val, success = success))
print(cat(paste("Iteration:", i, "\n"))) # Add iteration number
print(cat(paste("Spec: ", l, "\n")))
print(cat(paste("AIC: ", results_susini_oh$aic[i], "\n")))
print(cat(paste("Cox zph test statistic: ", results_susini_oh$cox_zph[i], "\n")))
}Iteration: 1
NULL
Spec: susinidum_oh *time
NULL
AIC: 239724.019577566
NULL
Cox zph test statistic: 4672.22758383257
NULL
Iteration: 2
NULL
Spec: susinidum_oh *log(time+0.001)
NULL
AIC: 241585.150871146
NULL
Cox zph test statistic: 1204.10177494143
NULL
Iteration: 3
NULL
Spec: susinidum_oh *poly(ifelse(is.na(time), .001,time), 3)
NULL
AIC: 240219.630415016
NULL
Cox zph test statistic: 7926.39955359493
NULL
Iteration: 4
NULL
Spec: susinidum_oh *poly(ifelse(is.na(time), .001,time), 4)
NULL
AIC: 240063.039259188
NULL
Cox zph test statistic: 8212.21889149851
NULL
Iteration: 5
NULL
Spec: susinidum_oh *poly(ifelse(is.na(time), .001,time), 5)
NULL
AIC: 240016.095427946
NULL
Cox zph test statistic: 8304.47329075385
NULL
Iteration: 6
NULL
Spec: susinidum_oh *poly(ifelse(is.na(time), .001,time), 6)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 7
NULL
Spec: susinidum_oh *pspline(log(time+0.001), df=3)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 8
NULL
Spec: susinidum_oh *pspline(log(time+0.001), df=4)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 9
NULL
Spec: susinidum_oh *pspline(log(time+0.001), df=5)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 10
NULL
Spec: susinidum_oh *pspline(log(time+0.001), df=6)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 11
NULL
Spec: susinidum_oh *bs(time, 3)
NULL
AIC: 239518.026724213
NULL
Cox zph test statistic: 10331.3482921498
NULL
Iteration: 12
NULL
Spec: susinidum_oh *bs(time, 4)
NULL
AIC: 236772.640457164
NULL
Cox zph test statistic: 18133.928243881
NULL
Iteration: 13
NULL
Spec: susinidum_oh *bs(time, 5)
NULL
AIC: 235970.965928735
NULL
Cox zph test statistic: 20900.8309746657
NULL
Iteration: 14
NULL
Spec: susinidum_oh *bs(time, 6)
NULL
AIC: 235677.262806318
NULL
Cox zph test statistic: 21727.8304904265
NULL
Iteration: 15
NULL
Spec: susinidum_oh *ns(time, 3)
NULL
AIC: 238253.396508616
NULL
Cox zph test statistic: 14728.7767838593
NULL
Iteration: 16
NULL
Spec: susinidum_oh *ns(time, 4)
NULL
AIC: 236653.226858842
NULL
Cox zph test statistic: 19308.1808541226
NULL
Iteration: 17
NULL
Spec: susinidum_oh *ns(time, 5)
NULL
AIC: 236251.069626771
NULL
Cox zph test statistic: 20993.031671471
NULL
Iteration: 18
NULL
Spec: susinidum_oh *ns(time, 6)
NULL
AIC: 235896.939555169
NULL
Cox zph test statistic: 21101.9090902985
NULL
Iteration: 19
NULL
Spec: susinidum_oh *rcs(time, 3)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULL
X matrix deemed to be singular; variable time''
Iteration: 20
NULL
Spec: susinidum_oh *rcs(time, 4)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULL
X matrix deemed to be singular; variable time'' time'''
Iteration: 21
NULL
Spec: susinidum_oh *rcs(time, 5)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULL
X matrix deemed to be singular; variable time'' time'''
Iteration: 22
NULL
Spec: susinidum_oh *rcs(time, 6)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULLCode
results_susini_oh[which(as.numeric(results_susini_oh$cox_zph)==
min(as.numeric(results_susini_oh$cox_zph))),] spec aic cox_zph success
2 susinidum_oh *log(time+0.001) 241585.2 1204.102 TRUECode
# spec aic cox_zph success
# 2 susinidum_oh *log(time+0.001) 241573 1203.69 TRUE
invisible("It worsen. We do nothing")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_susinidum_oh_alt <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq(1,5,.1)) {
change_points <- c(5*i, 10*i, 15*i) # Example change points (months)
coefficients <- c(0.5, 1.5, 0.5, 1) # Example coefficients
# Add the time-varying coefficient to your model
data_mine_miss_restr_proc2$susinidum_oh_rec2 <- with(data_mine_miss_restr_proc2,
susinidum_oh*time_varying_coeff(time, change_points, coefficients))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh_alt<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh_rec2 +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh_alt)
invisible("")
print(paste0("iteration: ",i))
aic_iter <- round(extractAIC(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh_alt)[[2]],1)
print( paste0("Time points: ",change_points[1],", ", change_points[2], ", ",change_points[3]))
print( paste0("AIC: ", aic_iter))
cox_zph_iter <- format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh_alt)$table), 1])
print(paste0("Cox ZPH: ", cox_zph_iter))
results_susinidum_oh_alt <- rbind.data.frame(results_susinidum_oh_alt,
cbind.data.frame(
spec = paste0("iteration: ",i, "; change points: ", change_points[1],", ", change_points[2], ", ",change_points[3]),
aic = aic_iter,
cox_zph = cox_zph_iter,
success = T
))
}[1] "iteration: 1"
[1] "Time points: 5, 10, 15"
[1] "AIC: 242702.9"
[1] "Cox ZPH: 345.489"
[1] "iteration: 1.1"
[1] "Time points: 5.5, 11, 16.5"
[1] "AIC: 242711.2"
[1] "Cox ZPH: 329.665"
[1] "iteration: 1.2"
[1] "Time points: 6, 12, 18"
[1] "AIC: 242707.8"
[1] "Cox ZPH: 329.167"
[1] "iteration: 1.3"
[1] "Time points: 6.5, 13, 19.5"
[1] "AIC: 242722"
[1] "Cox ZPH: 337.8074"
[1] "iteration: 1.4"
[1] "Time points: 7, 14, 21"
[1] "AIC: 242730.6"
[1] "Cox ZPH: 360.014"
[1] "iteration: 1.5"
[1] "Time points: 7.5, 15, 22.5"
[1] "AIC: 242736.7"
[1] "Cox ZPH: 396.3886"
[1] "iteration: 1.6"
[1] "Time points: 8, 16, 24"
[1] "AIC: 242739.4"
[1] "Cox ZPH: 415.7995"
[1] "iteration: 1.7"
[1] "Time points: 8.5, 17, 25.5"
[1] "AIC: 242741.2"
[1] "Cox ZPH: 449.8164"
[1] "iteration: 1.8"
[1] "Time points: 9, 18, 27"
[1] "AIC: 242740.8"
[1] "Cox ZPH: 465.6994"
[1] "iteration: 1.9"
[1] "Time points: 9.5, 19, 28.5"
[1] "AIC: 242739"
[1] "Cox ZPH: 484.3736"
[1] "iteration: 2"
[1] "Time points: 10, 20, 30"
[1] "AIC: 242735.2"
[1] "Cox ZPH: 495.5218"
[1] "iteration: 2.1"
[1] "Time points: 10.5, 21, 31.5"
[1] "AIC: 242732.2"
[1] "Cox ZPH: 468.4899"
[1] "iteration: 2.2"
[1] "Time points: 11, 22, 33"
[1] "AIC: 242722.2"
[1] "Cox ZPH: 451.4285"
[1] "iteration: 2.3"
[1] "Time points: 11.5, 23, 34.5"
[1] "AIC: 242718.2"
[1] "Cox ZPH: 423.5481"
[1] "iteration: 2.4"
[1] "Time points: 12, 24, 36"
[1] "AIC: 242716.3"
[1] "Cox ZPH: 395.6226"
[1] "iteration: 2.5"
[1] "Time points: 12.5, 25, 37.5"
[1] "AIC: 242716.8"
[1] "Cox ZPH: 364.4011"
[1] "iteration: 2.6"
[1] "Time points: 13, 26, 39"
[1] "AIC: 242717.5"
[1] "Cox ZPH: 340.7294"
[1] "iteration: 2.7"
[1] "Time points: 13.5, 27, 40.5"
[1] "AIC: 242720.7"
[1] "Cox ZPH: 331.1452"
[1] "iteration: 2.8"
[1] "Time points: 14, 28, 42"
[1] "AIC: 242724.9"
[1] "Cox ZPH: 334.6488"
[1] "iteration: 2.9"
[1] "Time points: 14.5, 29, 43.5"
[1] "AIC: 242729.4"
[1] "Cox ZPH: 353.4168"
[1] "iteration: 3"
[1] "Time points: 15, 30, 45"
[1] "AIC: 242733.9"
[1] "Cox ZPH: 395.1191"
[1] "iteration: 3.1"
[1] "Time points: 15.5, 31, 46.5"
[1] "AIC: 242737"
[1] "Cox ZPH: 455.8316"
[1] "iteration: 3.2"
[1] "Time points: 16, 32, 48"
[1] "AIC: 242740.4"
[1] "Cox ZPH: 524.5127"
[1] "iteration: 3.3"
[1] "Time points: 16.5, 33, 49.5"
[1] "AIC: 242741.2"
[1] "Cox ZPH: 601.5584"
[1] "iteration: 3.4"
[1] "Time points: 17, 34, 51"
[1] "AIC: 242740.6"
[1] "Cox ZPH: 686.4199"
[1] "iteration: 3.5"
[1] "Time points: 17.5, 35, 52.5"
[1] "AIC: 242738.5"
[1] "Cox ZPH: 771.4445"
[1] "iteration: 3.6"
[1] "Time points: 18, 36, 54"
[1] "AIC: 242733.6"
[1] "Cox ZPH: 868.9633"
[1] "iteration: 3.7"
[1] "Time points: 18.5, 37, 55.5"
[1] "AIC: 242730.4"
[1] "Cox ZPH: 951.8658"
[1] "iteration: 3.8"
[1] "Time points: 19, 38, 57"
[1] "AIC: 242721.3"
[1] "Cox ZPH: 1085.059"
[1] "iteration: 3.9"
[1] "Time points: 19.5, 39, 58.5"
[1] "AIC: 242709.4"
[1] "Cox ZPH: 1204.189"
[1] "iteration: 4"
[1] "Time points: 20, 40, 60"
[1] "AIC: 242693.7"
[1] "Cox ZPH: 1337.112"
[1] "iteration: 4.1"
[1] "Time points: 20.5, 41, 61.5"
[1] "AIC: 242682"
[1] "Cox ZPH: 1438.932"
[1] "iteration: 4.2"
[1] "Time points: 21, 42, 63"
[1] "AIC: 242664.9"
[1] "Cox ZPH: 1562.72"
[1] "iteration: 4.3"
[1] "Time points: 21.5, 43, 64.5"
[1] "AIC: 242651.1"
[1] "Cox ZPH: 1674.387"
[1] "iteration: 4.4"
[1] "Time points: 22, 44, 66"
[1] "AIC: 242620.8"
[1] "Cox ZPH: 1769.552"
[1] "iteration: 4.5"
[1] "Time points: 22.5, 45, 67.5"
[1] "AIC: 242607.7"
[1] "Cox ZPH: 1846.276"
[1] "iteration: 4.6"
[1] "Time points: 23, 46, 69"
[1] "AIC: 242585.1"
[1] "Cox ZPH: 1908.988"
[1] "iteration: 4.7"
[1] "Time points: 23.5, 47, 70.5"
[1] "AIC: 242569.1"
[1] "Cox ZPH: 1995.936"
[1] "iteration: 4.8"
[1] "Time points: 24, 48, 72"
[1] "AIC: 242551.2"
[1] "Cox ZPH: 2104.284"
[1] "iteration: 4.9"
[1] "Time points: 24.5, 49, 73.5"
[1] "AIC: 242534.8"
[1] "Cox ZPH: 2177.389"
[1] "iteration: 5"
[1] "Time points: 25, 50, 75"
[1] "AIC: 242517.7"
[1] "Cox ZPH: 2242.511"Code
results_susinidum_oh_alt[which(as.numeric(results_susinidum_oh_alt$cox_zph)==
min(as.numeric(results_susinidum_oh_alt$cox_zph))),] spec aic cox_zph success
3 iteration: 1.2; change points: 6, 12, 18 242707.8 329.167 TRUECode
# spec aic cox_zph success
# 3 iteration: 1.2; change points: 6, 12, 18 242696 328.8011 TRUE
message(paste0("The minimum of the first modified model: ",
format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)$table), 1]),
" vs. the minimum model: ", results_susinidum_oh_alt[which(as.numeric(results_susinidum_oh_alt$cox_zph)== min(as.numeric(results_susinidum_oh_alt$cox_zph))),"cox_zph"]
))Code
#The minimum of the first modified model: 321.3004 vs. the minimum model: 328.8011
invisible("It worsens. We do not integrate this model")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Analyse the violation of proportional hazards")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
try(dev.off())null device
1 Code
z_240422_5_2_restr <- loess(res240422_5_restr[,"susinidum_coc"] ~ time240422_5_restr, span=0.50) # residuals
par(mfrow= c(3,1))
plot(time240422_5_restr, fitted(z_240422_5_2_restr))
lines(supsmu(time240422_5_restr, res240422_5_restr[,"susinidum_coc"]),lty=2)
title("Sch. res., susinidum_coc")
plot(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)[which(attr(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out$coefficients, "names") == "susinidum_coc")], lwd=2, main= "Residuals of susinidum_coc")
abline(0,0, col="red", lty=3, lwd=2)
abline(h=model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out$coefficients["susinidum_coc"], col=3, lwd=2, lty= 3)
legend("bottomleft", legend= c("Reference line for the null effect", "Average hazard over time", "Time-varying hazard"), lty=c(3,2,1), col= c("red",3,1), lwd=2, bty = "n", bg = "transparent")
km_susinidum_coc <- survfit(Surv(lag_time,time,event)~ susinidum_coc+
cluster(id), data = data_mine_miss_restr_proc2)
plot(km_susinidum_coc, fun = "cloglog", xlab = "Time (in days) using log",col=c("black", "red"), ylab = "log-log survival", main = "log-log curves by susinidum_coc")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1, bty = "n", bg = "transparent")
recorded_plot <- recordPlot()
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
png(paste0(folder_path,"zph/proportionallity_susinidum_coc_restr.png"), height=13, width=10, res=500, units="in")
recorded_plot
dev.off()pdf
2 Code
invisible("Now with Initial substance: Cocaine")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_susinidum_coc <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq(.1,1.5,.025)) {
tvcs <- c(20,80)*i
data_mine_miss_restr_proc2$susinidum_coc_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$susinidum_coc,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc_rec +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc)
invisible("")
print(paste0("iteration: ",i))
aic_iter <- round(extractAIC(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc)[[2]],1)
print( paste0("Time points: ",tvcs[1],", ", tvcs[2]))
print( paste0("AIC: ", aic_iter))
cox_zph_iter <- format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc)$table), 1])
print(paste0("Cox ZPH: ", cox_zph_iter))
results_susinidum_coc <- rbind.data.frame(results_susinidum_coc,
cbind.data.frame(
spec = paste0("iteration: ",i, "; tvcs: ", tvcs[1],", ", tvcs[2]),
aic = aic_iter,
cox_zph = cox_zph_iter,
success = T
))
}[1] "iteration: 0.1"
[1] "Time points: 2, 8"
[1] "AIC: 242541.3"
[1] "Cox ZPH: 475.1056"
[1] "iteration: 0.125"
[1] "Time points: 2.5, 10"
[1] "AIC: 242534.1"
[1] "Cox ZPH: 489.6475"
[1] "iteration: 0.15"
[1] "Time points: 3, 12"
[1] "AIC: 242528.2"
[1] "Cox ZPH: 500.2812"
[1] "iteration: 0.175"
[1] "Time points: 3.5, 14"
[1] "AIC: 242524.4"
[1] "Cox ZPH: 504.4021"
[1] "iteration: 0.2"
[1] "Time points: 4, 16"
[1] "AIC: 242522.7"
[1] "Cox ZPH: 501.8106"
[1] "iteration: 0.225"
[1] "Time points: 4.5, 18"
[1] "AIC: 242523.2"
[1] "Cox ZPH: 492.8092"
[1] "iteration: 0.25"
[1] "Time points: 5, 20"
[1] "AIC: 242525.5"
[1] "Cox ZPH: 479.0295"
[1] "iteration: 0.275"
[1] "Time points: 5.5, 22"
[1] "AIC: 242530.1"
[1] "Cox ZPH: 458.4139"
[1] "iteration: 0.3"
[1] "Time points: 6, 24"
[1] "AIC: 242536.2"
[1] "Cox ZPH: 435.6086"
[1] "iteration: 0.325"
[1] "Time points: 6.5, 26"
[1] "AIC: 242543.6"
[1] "Cox ZPH: 411.411"
[1] "iteration: 0.35"
[1] "Time points: 7, 28"
[1] "AIC: 242553.4"
[1] "Cox ZPH: 384.4102"
[1] "iteration: 0.375"
[1] "Time points: 7.5, 30"
[1] "AIC: 242564.6"
[1] "Cox ZPH: 359.8189"
[1] "iteration: 0.4"
[1] "Time points: 8, 32"
[1] "AIC: 242576.1"
[1] "Cox ZPH: 342.9774"
[1] "iteration: 0.425"
[1] "Time points: 8.5, 34"
[1] "AIC: 242587.3"
[1] "Cox ZPH: 337.1026"
[1] "iteration: 0.45"
[1] "Time points: 9, 36"
[1] "AIC: 242597.7"
[1] "Cox ZPH: 343.4623"
[1] "iteration: 0.475"
[1] "Time points: 9.5, 38"
[1] "AIC: 242607.8"
[1] "Cox ZPH: 362.4096"
[1] "iteration: 0.5"
[1] "Time points: 10, 40"
[1] "AIC: 242617.6"
[1] "Cox ZPH: 394.2462"
[1] "iteration: 0.525"
[1] "Time points: 10.5, 42"
[1] "AIC: 242627.6"
[1] "Cox ZPH: 440.5531"
[1] "iteration: 0.55"
[1] "Time points: 11, 44"
[1] "AIC: 242638.2"
[1] "Cox ZPH: 503.2296"
[1] "iteration: 0.575"
[1] "Time points: 11.5, 46"
[1] "AIC: 242648.6"
[1] "Cox ZPH: 577.8865"
[1] "iteration: 0.6"
[1] "Time points: 12, 48"
[1] "AIC: 242659.1"
[1] "Cox ZPH: 662.8315"
[1] "iteration: 0.625"
[1] "Time points: 12.5, 50"
[1] "AIC: 242669.5"
[1] "Cox ZPH: 753.6261"
[1] "iteration: 0.65"
[1] "Time points: 13, 52"
[1] "AIC: 242679.9"
[1] "Cox ZPH: 846.3598"
[1] "iteration: 0.675"
[1] "Time points: 13.5, 54"
[1] "AIC: 242689.6"
[1] "Cox ZPH: 931.1736"
[1] "iteration: 0.7"
[1] "Time points: 14, 56"
[1] "AIC: 242698.7"
[1] "Cox ZPH: 1007.856"
[1] "iteration: 0.725"
[1] "Time points: 14.5, 58"
[1] "AIC: 242707.1"
[1] "Cox ZPH: 1071.06"
[1] "iteration: 0.75"
[1] "Time points: 15, 60"
[1] "AIC: 242715"
[1] "Cox ZPH: 1120.409"
[1] "iteration: 0.775"
[1] "Time points: 15.5, 62"
[1] "AIC: 242722.2"
[1] "Cox ZPH: 1154.777"
[1] "iteration: 0.8"
[1] "Time points: 16, 64"
[1] "AIC: 242728.3"
[1] "Cox ZPH: 1176.89"
[1] "iteration: 0.825"
[1] "Time points: 16.5, 66"
[1] "AIC: 242733.5"
[1] "Cox ZPH: 1187.06"
[1] "iteration: 0.85"
[1] "Time points: 17, 68"
[1] "AIC: 242737.5"
[1] "Cox ZPH: 1183.687"
[1] "iteration: 0.875"
[1] "Time points: 17.5, 70"
[1] "AIC: 242740.1"
[1] "Cox ZPH: 1169.481"
[1] "iteration: 0.9"
[1] "Time points: 18, 72"
[1] "AIC: 242741.3"
[1] "Cox ZPH: 1145.532"
[1] "iteration: 0.925"
[1] "Time points: 18.5, 74"
[1] "AIC: 242740.9"
[1] "Cox ZPH: 1112.816"
[1] "iteration: 0.95"
[1] "Time points: 19, 76"
[1] "AIC: 242738.7"
[1] "Cox ZPH: 1073.512"
[1] "iteration: 0.975"
[1] "Time points: 19.5, 78"
[1] "AIC: 242734.7"
[1] "Cox ZPH: 1029.364"
[1] "iteration: 1"
[1] "Time points: 20, 80"
[1] "AIC: 242728.6"
[1] "Cox ZPH: 981.5722"
[1] "iteration: 1.025"
[1] "Time points: 20.5, 82"
[1] "AIC: 242720.4"
[1] "Cox ZPH: 932.6265"
[1] "iteration: 1.05"
[1] "Time points: 21, 84"
[1] "AIC: 242709.7"
[1] "Cox ZPH: 881.6735"
[1] "iteration: 1.075"
[1] "Time points: 21.5, 86"
[1] "AIC: 242696.1"
[1] "Cox ZPH: 829.7161"
[1] "iteration: 1.1"
[1] "Time points: 22, 88"
[1] "AIC: 242679.4"
[1] "Cox ZPH: 776.9707"
[1] "iteration: 1.125"
[1] "Time points: 22.5, 90"
[1] "AIC: 242659.1"
[1] "Cox ZPH: 724.111"
[1] "iteration: 1.15"
[1] "Time points: 23, 92"
[1] "AIC: 242634.5"
[1] "Cox ZPH: 670.8339"
[1] "iteration: 1.175"
[1] "Time points: 23.5, 94"
[1] "AIC: 242605.3"
[1] "Cox ZPH: 617.9034"
[1] "iteration: 1.2"
[1] "Time points: 24, 96"
[1] "AIC: 242571.5"
[1] "Cox ZPH: 566.5202"
[1] "iteration: 1.225"
[1] "Time points: 24.5, 98"
[1] "AIC: 242534.1"
[1] "Cox ZPH: 518.4305"
[1] "iteration: 1.25"
[1] "Time points: 25, 100"
[1] "AIC: 242494.1"
[1] "Cox ZPH: 483.4053"
[1] "iteration: 1.275"
[1] "Time points: 25.5, 102"
[1] "AIC: 242453.5"
[1] "Cox ZPH: 449.0626"
[1] "iteration: 1.3"
[1] "Time points: 26, 104"
[1] "AIC: 242423.1"
[1] "Cox ZPH: 428.7996"
[1] "iteration: 1.325"
[1] "Time points: 26.5, 106"
[1] "AIC: 242394.9"
[1] "Cox ZPH: 415.4705"
[1] "iteration: 1.35"
[1] "Time points: 27, 108"
[1] "AIC: 242371.7"
[1] "Cox ZPH: 411.3971"
[1] "iteration: 1.375"
[1] "Time points: 27.5, 110"
[1] "AIC: 242350"
[1] "Cox ZPH: 424.4674"
[1] "iteration: 1.4"
[1] "Time points: 28, 112"
[1] "AIC: 242330.3"
[1] "Cox ZPH: 474.1608"
[1] "iteration: 1.425"
[1] "Time points: 28.5, 114"
[1] "AIC: 242313.6"
[1] "Cox ZPH: 581.5571"
[1] "iteration: 1.45"
[1] "Time points: 29, 116"
[1] "AIC: 242304.5"
[1] "Cox ZPH: 701.7825"
[1] "iteration: 1.475"
[1] "Time points: 29.5, 118"
[1] "AIC: 242301.8"
[1] "Cox ZPH: 757.7131"
[1] "iteration: 1.5"
[1] "Time points: 30, 120"
[1] "AIC: 242301.4"
[1] "Cox ZPH: 769.7678"Code
results_susinidum_coc[which(as.numeric(results_susinidum_coc$cox_zph)==
min(as.numeric(results_susinidum_coc$cox_zph))),] spec aic cox_zph success
14 iteration: 0.425; tvcs: 8.5, 34 242587.3 337.1026 TRUECode
# spec aic cox_zph success
# 14 iteration: 0.425; tvcs: 8.5, 34 242575.7 336.8893 TRUE
message(paste0("The previously modified model: ",
round(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)$table), 1],1),
" vs. the minimum model: ", results_susinidum_coc[which(as.numeric(results_susinidum_coc$cox_zph)== min(as.numeric(results_susinidum_coc$cox_zph))),"cox_zph"]
))Code
#The previously modified model: 321.3 vs. the minimum model: 336.8893
invisible("It worsen")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
require(splines)
specs_susinidum_coc <- paste0("susinidum_coc *", c("time", "log(time+0.001)", "poly(ifelse(is.na(time), .001,time), 3)", "poly(ifelse(is.na(time), .001,time), 4)", "poly(ifelse(is.na(time), .001,time), 5)", "poly(ifelse(is.na(time), .001,time), 6)", "pspline(log(time+0.001), df=3)", "pspline(log(time+0.001), df=4)", "pspline(log(time+0.001), df=5)", "pspline(log(time+0.001), df=6)", "bs(time, 3)", "bs(time, 4)", "bs(time, 5)", "bs(time, 6)", "ns(time, 3)", "ns(time, 4)", "ns(time, 5)", "ns(time, 6)", "rcs(time, 3)", "rcs(time, 4)", "rcs(time, 5)", "rcs(time, 6)"))
results_susini_coc <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq_along(specs_susinidum_coc)) {
l <- specs_susinidum_coc[i]
result <- tryCatch({
model <- cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
I(round(",l,",2)) +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
dplyr::mutate(log_lag_dias_treat_imp_sin_na_rec = bs(log_lag_dias_treat_imp_sin_na, 3)) %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
aic_val <- extractAIC(model)[[2]]
cox_zph_val <- cox.zph(model)$table[nrow(cox.zph(model)$table), 1]
success <- TRUE
}, error = function(e) {
aic_val <- NA
cox_zph_val <- NA
success <- FALSE
warning(paste0("Error fitting specification ", i, " (", l, "): ", e))
})
results_susini_coc <- rbind(results_susini_coc, data.frame(spec = l, aic = aic_val, cox_zph = cox_zph_val, success = success))
print(cat(paste("Iteration:", i, "\n"))) # Add iteration number
print(cat(paste("Spec: ", l, "\n")))
print(cat(paste("AIC: ", results_susini_oh$aic[i], "\n")))
print(cat(paste("Cox zph test statistic: ", results_susini_oh$cox_zph[i], "\n")))
}Iteration: 1
NULL
Spec: susinidum_coc *time
NULL
AIC: 239724.019577566
NULL
Cox zph test statistic: 4672.22758383257
NULL
Iteration: 2
NULL
Spec: susinidum_coc *log(time+0.001)
NULL
AIC: 241585.150871146
NULL
Cox zph test statistic: 1204.10177494143
NULL
Iteration: 3
NULL
Spec: susinidum_coc *poly(ifelse(is.na(time), .001,time), 3)
NULL
AIC: 240219.630415016
NULL
Cox zph test statistic: 7926.39955359493
NULL
Iteration: 4
NULL
Spec: susinidum_coc *poly(ifelse(is.na(time), .001,time), 4)
NULL
AIC: 240063.039259188
NULL
Cox zph test statistic: 8212.21889149851
NULL
Iteration: 5
NULL
Spec: susinidum_coc *poly(ifelse(is.na(time), .001,time), 5)
NULL
AIC: 240016.095427946
NULL
Cox zph test statistic: 8304.47329075385
NULL
Iteration: 6
NULL
Spec: susinidum_coc *poly(ifelse(is.na(time), .001,time), 6)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 7
NULL
Spec: susinidum_coc *pspline(log(time+0.001), df=3)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 8
NULL
Spec: susinidum_coc *pspline(log(time+0.001), df=4)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 9
NULL
Spec: susinidum_coc *pspline(log(time+0.001), df=5)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 10
NULL
Spec: susinidum_coc *pspline(log(time+0.001), df=6)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 11
NULL
Spec: susinidum_coc *bs(time, 3)
NULL
AIC: 239518.026724213
NULL
Cox zph test statistic: 10331.3482921498
NULL
Iteration: 12
NULL
Spec: susinidum_coc *bs(time, 4)
NULL
AIC: 236772.640457164
NULL
Cox zph test statistic: 18133.928243881
NULL
Iteration: 13
NULL
Spec: susinidum_coc *bs(time, 5)
NULL
AIC: 235970.965928735
NULL
Cox zph test statistic: 20900.8309746657
NULL
Iteration: 14
NULL
Spec: susinidum_coc *bs(time, 6)
NULL
AIC: 235677.262806318
NULL
Cox zph test statistic: 21727.8304904265
NULL
Iteration: 15
NULL
Spec: susinidum_coc *ns(time, 3)
NULL
AIC: 238253.396508616
NULL
Cox zph test statistic: 14728.7767838593
NULL
Iteration: 16
NULL
Spec: susinidum_coc *ns(time, 4)
NULL
AIC: 236653.226858842
NULL
Cox zph test statistic: 19308.1808541226
NULL
Iteration: 17
NULL
Spec: susinidum_coc *ns(time, 5)
NULL
AIC: 236251.069626771
NULL
Cox zph test statistic: 20993.031671471
NULL
Iteration: 18
NULL
Spec: susinidum_coc *ns(time, 6)
NULL
AIC: 235896.939555169
NULL
Cox zph test statistic: 21101.9090902985
NULL
Iteration: 19
NULL
Spec: susinidum_coc *rcs(time, 3)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULL
Iteration: 20
NULL
Spec: susinidum_coc *rcs(time, 4)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULL
Iteration: 21
NULL
Spec: susinidum_coc *rcs(time, 5)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULL
X matrix deemed to be singular; variable time'''
Iteration: 22
NULL
Spec: susinidum_coc *rcs(time, 6)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULLCode
results_susini_coc[which(as.numeric(results_susini_coc$cox_zph)==
min(as.numeric(results_susini_coc$cox_zph))),] spec aic cox_zph success
2 susinidum_coc *log(time+0.001) 242707.4 323.3394 TRUECode
# spec aic cox_zph success
# 2 susinidum_coc *log(time+0.001) 242695.6 323.038 TRUE
invisible("It worsen. We do nothing")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_susinidum_coc_alt <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq(1,5,.1)) {
change_points <- c(5*i, 10*i, 15*i) # Example change points (months)
coefficients <- c(0.5, 1.5, 0.5, 1) # Example coefficients
# Add the time-varying coefficient to your model
data_mine_miss_restr_proc2$susinidum_coc_rec2 <- with(data_mine_miss_restr_proc2,
susinidum_coc*time_varying_coeff(time, change_points, coefficients))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)
invisible("")
print(paste0("iteration: ",i))
aic_iter <- round(extractAIC(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)[[2]],1)
print( paste0("Time points: ",change_points[1],", ", change_points[2], ", ",change_points[3]))
print( paste0("AIC: ", aic_iter))
cox_zph_iter <- format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)$table), 1])
print(paste0("Cox ZPH: ", cox_zph_iter))
results_susinidum_coc_alt <- rbind.data.frame(results_susinidum_coc_alt,
cbind.data.frame(
spec = paste0("iteration: ",i, "; change points: ", change_points[1],", ", change_points[2], ", ",change_points[3]),
aic = aic_iter,
cox_zph = cox_zph_iter,
success = T
))
}[1] "iteration: 1"
[1] "Time points: 5, 10, 15"
[1] "AIC: 242740.9"
[1] "Cox ZPH: 318.8496"
[1] "iteration: 1.1"
[1] "Time points: 5.5, 11, 16.5"
[1] "AIC: 242740.9"
[1] "Cox ZPH: 319.5982"
[1] "iteration: 1.2"
[1] "Time points: 6, 12, 18"
[1] "AIC: 242741"
[1] "Cox ZPH: 321.6762"
[1] "iteration: 1.3"
[1] "Time points: 6.5, 13, 19.5"
[1] "AIC: 242740.9"
[1] "Cox ZPH: 321.3254"
[1] "iteration: 1.4"
[1] "Time points: 7, 14, 21"
[1] "AIC: 242740.8"
[1] "Cox ZPH: 321.5408"
[1] "iteration: 1.5"
[1] "Time points: 7.5, 15, 22.5"
[1] "AIC: 242740.6"
[1] "Cox ZPH: 322.0213"
[1] "iteration: 1.6"
[1] "Time points: 8, 16, 24"
[1] "AIC: 242741"
[1] "Cox ZPH: 324.2359"
[1] "iteration: 1.7"
[1] "Time points: 8.5, 17, 25.5"
[1] "AIC: 242741.3"
[1] "Cox ZPH: 327.4271"
[1] "iteration: 1.8"
[1] "Time points: 9, 18, 27"
[1] "AIC: 242741.3"
[1] "Cox ZPH: 327.9855"
[1] "iteration: 1.9"
[1] "Time points: 9.5, 19, 28.5"
[1] "AIC: 242741.3"
[1] "Cox ZPH: 328.6902"
[1] "iteration: 2"
[1] "Time points: 10, 20, 30"
[1] "AIC: 242741.1"
[1] "Cox ZPH: 329.7676"
[1] "iteration: 2.1"
[1] "Time points: 10.5, 21, 31.5"
[1] "AIC: 242741"
[1] "Cox ZPH: 327.3083"
[1] "iteration: 2.2"
[1] "Time points: 11, 22, 33"
[1] "AIC: 242741"
[1] "Cox ZPH: 325.072"
[1] "iteration: 2.3"
[1] "Time points: 11.5, 23, 34.5"
[1] "AIC: 242740.3"
[1] "Cox ZPH: 324.9401"
[1] "iteration: 2.4"
[1] "Time points: 12, 24, 36"
[1] "AIC: 242739.6"
[1] "Cox ZPH: 323.5421"
[1] "iteration: 2.5"
[1] "Time points: 12.5, 25, 37.5"
[1] "AIC: 242740.1"
[1] "Cox ZPH: 321.4559"
[1] "iteration: 2.6"
[1] "Time points: 13, 26, 39"
[1] "AIC: 242739.5"
[1] "Cox ZPH: 319.2246"
[1] "iteration: 2.7"
[1] "Time points: 13.5, 27, 40.5"
[1] "AIC: 242739"
[1] "Cox ZPH: 318.7078"
[1] "iteration: 2.8"
[1] "Time points: 14, 28, 42"
[1] "AIC: 242738.8"
[1] "Cox ZPH: 317.9917"
[1] "iteration: 2.9"
[1] "Time points: 14.5, 29, 43.5"
[1] "AIC: 242739.3"
[1] "Cox ZPH: 317.7749"
[1] "iteration: 3"
[1] "Time points: 15, 30, 45"
[1] "AIC: 242740.1"
[1] "Cox ZPH: 317.7125"
[1] "iteration: 3.1"
[1] "Time points: 15.5, 31, 46.5"
[1] "AIC: 242740.4"
[1] "Cox ZPH: 317.9962"
[1] "iteration: 3.2"
[1] "Time points: 16, 32, 48"
[1] "AIC: 242741.2"
[1] "Cox ZPH: 319.1036"
[1] "iteration: 3.3"
[1] "Time points: 16.5, 33, 49.5"
[1] "AIC: 242741.3"
[1] "Cox ZPH: 320.6839"
[1] "iteration: 3.4"
[1] "Time points: 17, 34, 51"
[1] "AIC: 242741.3"
[1] "Cox ZPH: 323.1721"
[1] "iteration: 3.5"
[1] "Time points: 17.5, 35, 52.5"
[1] "AIC: 242740.6"
[1] "Cox ZPH: 326.611"
[1] "iteration: 3.6"
[1] "Time points: 18, 36, 54"
[1] "AIC: 242740.6"
[1] "Cox ZPH: 328.6784"
[1] "iteration: 3.7"
[1] "Time points: 18.5, 37, 55.5"
[1] "AIC: 242740"
[1] "Cox ZPH: 333.0428"
[1] "iteration: 3.8"
[1] "Time points: 19, 38, 57"
[1] "AIC: 242739.3"
[1] "Cox ZPH: 336.2208"
[1] "iteration: 3.9"
[1] "Time points: 19.5, 39, 58.5"
[1] "AIC: 242738.4"
[1] "Cox ZPH: 339.4407"
[1] "iteration: 4"
[1] "Time points: 20, 40, 60"
[1] "AIC: 242737.2"
[1] "Cox ZPH: 344.0601"
[1] "iteration: 4.1"
[1] "Time points: 20.5, 41, 61.5"
[1] "AIC: 242735.7"
[1] "Cox ZPH: 347.9692"
[1] "iteration: 4.2"
[1] "Time points: 21, 42, 63"
[1] "AIC: 242735"
[1] "Cox ZPH: 352.4328"
[1] "iteration: 4.3"
[1] "Time points: 21.5, 43, 64.5"
[1] "AIC: 242734.3"
[1] "Cox ZPH: 357.9312"
[1] "iteration: 4.4"
[1] "Time points: 22, 44, 66"
[1] "AIC: 242732.4"
[1] "Cox ZPH: 359.4456"
[1] "iteration: 4.5"
[1] "Time points: 22.5, 45, 67.5"
[1] "AIC: 242731"
[1] "Cox ZPH: 364.9348"
[1] "iteration: 4.6"
[1] "Time points: 23, 46, 69"
[1] "AIC: 242730.3"
[1] "Cox ZPH: 365.0121"
[1] "iteration: 4.7"
[1] "Time points: 23.5, 47, 70.5"
[1] "AIC: 242729.2"
[1] "Cox ZPH: 363.3641"
[1] "iteration: 4.8"
[1] "Time points: 24, 48, 72"
[1] "AIC: 242727.3"
[1] "Cox ZPH: 367.802"
[1] "iteration: 4.9"
[1] "Time points: 24.5, 49, 73.5"
[1] "AIC: 242728.2"
[1] "Cox ZPH: 370.3726"
[1] "iteration: 5"
[1] "Time points: 25, 50, 75"
[1] "AIC: 242727.4"
[1] "Cox ZPH: 370.8872"Code
results_susinidum_coc_alt[which(as.numeric(results_susinidum_coc_alt$cox_zph)==
min(as.numeric(results_susinidum_coc_alt$cox_zph))),] spec aic cox_zph success
21 iteration: 3; change points: 15, 30, 45 242740.1 317.7125 TRUECode
# spec aic cox_zph success
# 21 iteration: 3; change points: 15, 30, 45 242728.3 317.3692 TRUE
message(paste0("The minimum of the first modified model: ",
format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out)$table), 1]),
" vs. the minimum model: ", results_susinidum_coc_alt[which(as.numeric(results_susinidum_coc_alt$cox_zph)== min(as.numeric(results_susinidum_coc_alt$cox_zph))),"cox_zph"]
))Code
#The minimum of the first modified model: 321.3004 vs. the minimum model: 317.3692
invisible("It improves")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_susinidum_coc_alt[which(as.numeric(results_susinidum_coc_alt$cox_zph)==
min(as.numeric(results_susinidum_coc_alt$cox_zph))),] spec aic cox_zph success
21 iteration: 3; change points: 15, 30, 45 242740.1 317.7125 TRUECode
change_points <- c(5*3, 10*3, 15*3) # Example change points (months)
coefficients <- c(0.5, 1.5, 0.5, 1) # Example coefficients
# Add the time-varying coefficient to your model
data_mine_miss_restr_proc2$susinidum_coc_rec2 <- with(data_mine_miss_restr_proc2,
susinidum_coc*time_varying_coeff(time, change_points, coefficients))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt<-
cph(Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
round(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)$table,4) chisq df p
lag_tr_outcome_rec 66.6088 1 0.0000
log_lag_dias_treat_imp_sin_na 10.5341 1 0.0012
lag_less_90d_tr1_rec 84.1713 1 0.0000
lag_comp_bpsc_y3_severe_rec 17.3155 1 0.0000
lag_policonsumo2 3.2002 1 0.0736
edad_al_ing_1 0.1512 1 0.6974
ano_nac_corr 5.1470 1 0.0233
susinidum_coc_rec2 2.9483 1 0.0860
susinidum_oh 19.9650 1 0.0000
susinidum_pbc 2.0948 1 0.1478
susinidum_mar 6.7931 1 0.0092
psycom_dum_with 11.7373 1 0.0006
psycom_dum_study 1.6233 1 0.2026
freq_cons_dum_5day 0.0825 1 0.7739
cond_oc_dum_2inact 4.4758 1 0.0344
cond_oc_dum_3unemp 1.7447 1 0.1865
susprindum_oh 4.3698 1 0.0366
susprindum_coc 1.5662 1 0.2108
susprindum_pbc 0.0503 1 0.8226
susprindum_mar 5.2711 1 0.0217
GLOBAL 317.7125 20 0.0000Code
# chisq df p
# lag_tr_outcome_rec 66.7157 1 0.0000
# log_lag_dias_treat_imp_sin_na 10.3761 1 0.0013
# lag_less_90d_tr1_rec 84.1743 1 0.0000
# lag_comp_bpsc_y3_severe_rec 17.2693 1 0.0000
# lag_policonsumo2 3.1689 1 0.0751
# edad_al_ing_1 0.1432 1 0.7051
# ano_nac_corr 5.1839 1 0.0228
# susinidum_coc_rec2 2.9377 1 0.0865
# susinidum_oh 19.7620 1 0.0000
# susinidum_pbc 2.0158 1 0.1557
# susinidum_mar 6.7590 1 0.0093
# psycom_dum_with 11.7916 1 0.0006
# psycom_dum_study 1.6349 1 0.2010
# freq_cons_dum_5day 0.0541 1 0.8161
# cond_oc_dum_2inact 4.5002 1 0.0339
# cond_oc_dum_3unemp 1.6255 1 0.2023
# susprindum_oh 4.2826 1 0.0385
# susprindum_coc 1.6036 1 0.2054
# susprindum_pbc 0.0666 1 0.7964
# susprindum_mar 5.3207 1 0.0211
# GLOBAL 317.3692 20 0.0000
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Analyse the violation of proportional hazards")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
try(dev.off())null device
1 Code
res240422_6_restr <- resid(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt, "scaledsch")
time240422_6_restr <- as.numeric(dimnames(res240422_6_restr)[[1]])
z_240422_6_1_restr <- loess(res240422_6_restr[,"susinidum_oh"] ~ time240422_6_restr, span=0.50) # residuals
par(mfrow= c(3,1))
plot(time240422_6_restr, fitted(z_240422_6_1_restr))
lines(supsmu(time240422_6_restr, res240422_6_restr[,"susinidum_oh"]),lty=2)
title("Sch. res., susinidum_coc")
plot(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)[which(attr(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt$coefficients, "names") == "susinidum_oh")], lwd=2, main= "Residuals of susinidum_oh")
abline(0,0, col="red", lty=3, lwd=2)
abline(h=model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt$coefficients["susinidum_oh"], col=3, lwd=2, lty= 3)
legend("bottomleft", legend= c("Reference line for the null effect", "Average hazard over time", "Time-varying hazard"), lty=c(3,2,1), col= c("red",3,1), lwd=2, bty = "n", bg = "transparent")
km_susinidum_oh <- survfit(Surv(lag_time,time,event)~ susinidum_oh+
cluster(id), data = data_mine_miss_restr_proc2)
plot(km_susinidum_oh, fun = "cloglog", xlab = "Time (in days) using log",col=c("black", "red"), ylab = "log-log survival", main = "log-log curves by susinidum_oh")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1, bty = "n", bg = "transparent")
recorded_plot <- recordPlot()
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
png(paste0(folder_path,"zph/proportionallity_susinidum_oh2_restr.png"), height=13, width=10, res=500, units="in")
recorded_plot
dev.off()pdf
2 Code
invisible("Now with Initial substance: Alcohol AGAIN")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_susinidum_oh2 <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq(.1,1.5,.025)) {
tvcs <- c(20,80)*i
data_mine_miss_restr_proc2$susinidum_oh_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$susinidum_oh,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_susinidum_oh<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh_rec +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_susinidum_oh)
invisible("")
print(paste0("iteration: ",i))
aic_iter <- round(extractAIC(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_susinidum_oh)[[2]],1)
print( paste0("Time points: ",tvcs[1],", ", tvcs[2]))
print( paste0("AIC: ", aic_iter))
cox_zph_iter <- format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_susinidum_oh)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_susinidum_oh)$table), 1])
print(paste0("Cox ZPH: ", cox_zph_iter))
results_susinidum_oh2 <- rbind.data.frame(results_susinidum_oh2,
cbind.data.frame(
spec = paste0("iteration: ",i, "; tvcs: ", tvcs[1],", ", tvcs[2]),
aic = aic_iter,
cox_zph = cox_zph_iter,
success = T
))
}[1] "iteration: 0.1"
[1] "Time points: 2, 8"
[1] "AIC: 239556.7"
[1] "Cox ZPH: 5469.566"
[1] "iteration: 0.125"
[1] "Time points: 2.5, 10"
[1] "AIC: 239513.6"
[1] "Cox ZPH: 5660.287"
[1] "iteration: 0.15"
[1] "Time points: 3, 12"
[1] "AIC: 239499.1"
[1] "Cox ZPH: 5766.978"
[1] "iteration: 0.175"
[1] "Time points: 3.5, 14"
[1] "AIC: 239512.2"
[1] "Cox ZPH: 5799.257"
[1] "iteration: 0.2"
[1] "Time points: 4, 16"
[1] "AIC: 239556.4"
[1] "Cox ZPH: 5748.393"
[1] "iteration: 0.225"
[1] "Time points: 4.5, 18"
[1] "AIC: 239623.7"
[1] "Cox ZPH: 5636.161"
[1] "iteration: 0.25"
[1] "Time points: 5, 20"
[1] "AIC: 239711.5"
[1] "Cox ZPH: 5463.947"
[1] "iteration: 0.275"
[1] "Time points: 5.5, 22"
[1] "AIC: 239820.5"
[1] "Cox ZPH: 5222.299"
[1] "iteration: 0.3"
[1] "Time points: 6, 24"
[1] "AIC: 239953.8"
[1] "Cox ZPH: 4892.01"
[1] "iteration: 0.325"
[1] "Time points: 6.5, 26"
[1] "AIC: 240100.8"
[1] "Cox ZPH: 4497.324"
[1] "iteration: 0.35"
[1] "Time points: 7, 28"
[1] "AIC: 240262.6"
[1] "Cox ZPH: 4030.255"
[1] "iteration: 0.375"
[1] "Time points: 7.5, 30"
[1] "AIC: 240431.9"
[1] "Cox ZPH: 3523.814"
[1] "iteration: 0.4"
[1] "Time points: 8, 32"
[1] "AIC: 240604.4"
[1] "Cox ZPH: 3007.152"
[1] "iteration: 0.425"
[1] "Time points: 8.5, 34"
[1] "AIC: 240780.5"
[1] "Cox ZPH: 2499.49"
[1] "iteration: 0.45"
[1] "Time points: 9, 36"
[1] "AIC: 240959.3"
[1] "Cox ZPH: 2042.524"
[1] "iteration: 0.475"
[1] "Time points: 9.5, 38"
[1] "AIC: 241138"
[1] "Cox ZPH: 1695.105"
[1] "iteration: 0.5"
[1] "Time points: 10, 40"
[1] "AIC: 241309.6"
[1] "Cox ZPH: 1516.189"
[1] "iteration: 0.525"
[1] "Time points: 10.5, 42"
[1] "AIC: 241477.7"
[1] "Cox ZPH: 1530.988"
[1] "iteration: 0.55"
[1] "Time points: 11, 44"
[1] "AIC: 241642.2"
[1] "Cox ZPH: 1754.078"
[1] "iteration: 0.575"
[1] "Time points: 11.5, 46"
[1] "AIC: 241800.2"
[1] "Cox ZPH: 2173.52"
[1] "iteration: 0.6"
[1] "Time points: 12, 48"
[1] "AIC: 241950.1"
[1] "Cox ZPH: 2751.925"
[1] "iteration: 0.625"
[1] "Time points: 12.5, 50"
[1] "AIC: 242092.6"
[1] "Cox ZPH: 3446.823"
[1] "iteration: 0.65"
[1] "Time points: 13, 52"
[1] "AIC: 242227"
[1] "Cox ZPH: 4199.34"
[1] "iteration: 0.675"
[1] "Time points: 13.5, 54"
[1] "AIC: 242350.7"
[1] "Cox ZPH: 4944.467"
[1] "iteration: 0.7"
[1] "Time points: 14, 56"
[1] "AIC: 242460.7"
[1] "Cox ZPH: 5620.455"
[1] "iteration: 0.725"
[1] "Time points: 14.5, 58"
[1] "AIC: 242554.3"
[1] "Cox ZPH: 6178.287"
[1] "iteration: 0.75"
[1] "Time points: 15, 60"
[1] "AIC: 242630.7"
[1] "Cox ZPH: 6598.576"
[1] "iteration: 0.775"
[1] "Time points: 15.5, 62"
[1] "AIC: 242688.3"
[1] "Cox ZPH: 6868.943"
[1] "iteration: 0.8"
[1] "Time points: 16, 64"
[1] "AIC: 242725.3"
[1] "Cox ZPH: 6996.218"
[1] "iteration: 0.825"
[1] "Time points: 16.5, 66"
[1] "AIC: 242740.2"
[1] "Cox ZPH: 7000.32"
[1] "iteration: 0.85"
[1] "Time points: 17, 68"
[1] "AIC: 242732.2"
[1] "Cox ZPH: 6903.379"
[1] "iteration: 0.875"
[1] "Time points: 17.5, 70"
[1] "AIC: 242700"
[1] "Cox ZPH: 6727.182"
[1] "iteration: 0.9"
[1] "Time points: 18, 72"
[1] "AIC: 242643.3"
[1] "Cox ZPH: 6493.766"
[1] "iteration: 0.925"
[1] "Time points: 18.5, 74"
[1] "AIC: 242562.9"
[1] "Cox ZPH: 6225.363"
[1] "iteration: 0.95"
[1] "Time points: 19, 76"
[1] "AIC: 242458.8"
[1] "Cox ZPH: 5933.628"
[1] "iteration: 0.975"
[1] "Time points: 19.5, 78"
[1] "AIC: 242330.8"
[1] "Cox ZPH: 5629.225"
[1] "iteration: 1"
[1] "Time points: 20, 80"
[1] "AIC: 242178"
[1] "Cox ZPH: 5315.827"
[1] "iteration: 1.025"
[1] "Time points: 20.5, 82"
[1] "AIC: 242001.4"
[1] "Cox ZPH: 5006.006"
[1] "iteration: 1.05"
[1] "Time points: 21, 84"
[1] "AIC: 241803.4"
[1] "Cox ZPH: 4708.628"
[1] "iteration: 1.075"
[1] "Time points: 21.5, 86"
[1] "AIC: 241588.2"
[1] "Cox ZPH: 4429.636"
[1] "iteration: 1.1"
[1] "Time points: 22, 88"
[1] "AIC: 241355.7"
[1] "Cox ZPH: 4166.516"
[1] "iteration: 1.125"
[1] "Time points: 22.5, 90"
[1] "AIC: 241106.1"
[1] "Cox ZPH: 3917.886"
[1] "iteration: 1.15"
[1] "Time points: 23, 92"
[1] "AIC: 240845.4"
[1] "Cox ZPH: 3688.558"
[1] "iteration: 1.175"
[1] "Time points: 23.5, 94"
[1] "AIC: 240574.7"
[1] "Cox ZPH: 3476.06"
[1] "iteration: 1.2"
[1] "Time points: 24, 96"
[1] "AIC: 240294.8"
[1] "Cox ZPH: 3285.221"
[1] "iteration: 1.225"
[1] "Time points: 24.5, 98"
[1] "AIC: 240013.4"
[1] "Cox ZPH: 3118.936"
[1] "iteration: 1.25"
[1] "Time points: 25, 100"
[1] "AIC: 239734.8"
[1] "Cox ZPH: 2977.732"
[1] "iteration: 1.275"
[1] "Time points: 25.5, 102"
[1] "AIC: 239459.9"
[1] "Cox ZPH: 2862.778"
[1] "iteration: 1.3"
[1] "Time points: 26, 104"
[1] "AIC: 239188.1"
[1] "Cox ZPH: 2772.228"
[1] "iteration: 1.325"
[1] "Time points: 26.5, 106"
[1] "AIC: 238923.8"
[1] "Cox ZPH: 2710.248"
[1] "iteration: 1.35"
[1] "Time points: 27, 108"
[1] "AIC: 238664.5"
[1] "Cox ZPH: 2681.074"
[1] "iteration: 1.375"
[1] "Time points: 27.5, 110"
[1] "AIC: 238411.7"
[1] "Cox ZPH: 2689.043"
[1] "iteration: 1.4"
[1] "Time points: 28, 112"
[1] "AIC: 238178.4"
[1] "Cox ZPH: 2739.046"
[1] "iteration: 1.425"
[1] "Time points: 28.5, 114"
[1] "AIC: 237964.8"
[1] "Cox ZPH: 2841.23"
[1] "iteration: 1.45"
[1] "Time points: 29, 116"
[1] "AIC: 237767.9"
[1] "Cox ZPH: 3011.975"
[1] "iteration: 1.475"
[1] "Time points: 29.5, 118"
[1] "AIC: 237594.5"
[1] "Cox ZPH: 3274.361"
[1] "iteration: 1.5"
[1] "Time points: 30, 120"
[1] "AIC: 237451.1"
[1] "Cox ZPH: 3701.742"Code
results_susinidum_oh2[which(as.numeric(results_susinidum_oh2$cox_zph)==
min(as.numeric(results_susinidum_oh2$cox_zph))),] spec aic cox_zph success
17 iteration: 0.5; tvcs: 10, 40 241309.6 1516.189 TRUECode
# spec aic cox_zph success
# 15 iteration: 0.45; tvcs: 9, 36 242729.5 321.3004 TRUE
results_susinidum_coc_alt[which(as.numeric(results_susinidum_coc_alt$cox_zph)==
min(as.numeric(results_susinidum_coc_alt$cox_zph))),"cox_zph"][1] "317.7125"Code
message(paste0("The previously modified model: ",
round(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)$table), 1],1),
" vs. the minimum model: ", results_susinidum_oh2[which(as.numeric(results_susinidum_oh2$cox_zph)== min(as.numeric(results_susinidum_oh2$cox_zph))),"cox_zph"]
))Code
#The previously modified model: 326.3 vs. the minimum model: 1514.375
invisible("It worsen")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
require(splines)
specs_susinidum_oh <- paste0("susinidum_oh *", c("time", "log(time+0.001)", "poly(ifelse(is.na(time), .001,time), 3)", "poly(ifelse(is.na(time), .001,time), 4)", "poly(ifelse(is.na(time), .001,time), 5)", "poly(ifelse(is.na(time), .001,time), 6)", "pspline(log(time+0.001), df=3)", "pspline(log(time+0.001), df=4)", "pspline(log(time+0.001), df=5)", "pspline(log(time+0.001), df=6)", "bs(time, 3)", "bs(time, 4)", "bs(time, 5)", "bs(time, 6)", "ns(time, 3)", "ns(time, 4)", "ns(time, 5)", "ns(time, 6)", "rcs(time, 3)", "rcs(time, 4)", "rcs(time, 5)", "rcs(time, 6)"))
results_susini_oh2 <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq_along(specs_susinidum_oh)) {
l <- specs_susinidum_oh[i]
result <- tryCatch({
model <- cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
I(round(",l,",2)) +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
dplyr::mutate(log_lag_dias_treat_imp_sin_na_rec = bs(log_lag_dias_treat_imp_sin_na, 3)) %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
aic_val <- extractAIC(model)[[2]]
cox_zph_val <- cox.zph(model)$table[nrow(cox.zph(model)$table), 1]
success <- TRUE
}, error = function(e) {
aic_val <- NA
cox_zph_val <- NA
success <- FALSE
warning(paste0("Error fitting specification ", i, " (", l, "): ", e))
})
results_susini_oh2 <- rbind(results_susini_oh2, data.frame(spec = l, aic = aic_val, cox_zph = cox_zph_val, success = success))
print(cat(paste("Iteration:", i, "\n"))) # Add iteration number
print(cat(paste("Spec: ", l, "\n")))
print(cat(paste("AIC: ", results_susini_oh$aic[i], "\n")))
print(cat(paste("Cox zph test statistic: ", results_susini_oh$cox_zph[i], "\n")))
}Iteration: 1
NULL
Spec: susinidum_oh *time
NULL
AIC: 239724.019577566
NULL
Cox zph test statistic: 4672.22758383257
NULL
Iteration: 2
NULL
Spec: susinidum_oh *log(time+0.001)
NULL
AIC: 241585.150871146
NULL
Cox zph test statistic: 1204.10177494143
NULL
Iteration: 3
NULL
Spec: susinidum_oh *poly(ifelse(is.na(time), .001,time), 3)
NULL
AIC: 240219.630415016
NULL
Cox zph test statistic: 7926.39955359493
NULL
Iteration: 4
NULL
Spec: susinidum_oh *poly(ifelse(is.na(time), .001,time), 4)
NULL
AIC: 240063.039259188
NULL
Cox zph test statistic: 8212.21889149851
NULL
Iteration: 5
NULL
Spec: susinidum_oh *poly(ifelse(is.na(time), .001,time), 5)
NULL
AIC: 240016.095427946
NULL
Cox zph test statistic: 8304.47329075385
NULL
Iteration: 6
NULL
Spec: susinidum_oh *poly(ifelse(is.na(time), .001,time), 6)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 7
NULL
Spec: susinidum_oh *pspline(log(time+0.001), df=3)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 8
NULL
Spec: susinidum_oh *pspline(log(time+0.001), df=4)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 9
NULL
Spec: susinidum_oh *pspline(log(time+0.001), df=5)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 10
NULL
Spec: susinidum_oh *pspline(log(time+0.001), df=6)
NULL
AIC: 239508.178249717
NULL
Cox zph test statistic: 8101.01468172362
NULL
Iteration: 11
NULL
Spec: susinidum_oh *bs(time, 3)
NULL
AIC: 239518.026724213
NULL
Cox zph test statistic: 10331.3482921498
NULL
Iteration: 12
NULL
Spec: susinidum_oh *bs(time, 4)
NULL
AIC: 236772.640457164
NULL
Cox zph test statistic: 18133.928243881
NULL
Iteration: 13
NULL
Spec: susinidum_oh *bs(time, 5)
NULL
AIC: 235970.965928735
NULL
Cox zph test statistic: 20900.8309746657
NULL
Iteration: 14
NULL
Spec: susinidum_oh *bs(time, 6)
NULL
AIC: 235677.262806318
NULL
Cox zph test statistic: 21727.8304904265
NULL
Iteration: 15
NULL
Spec: susinidum_oh *ns(time, 3)
NULL
AIC: 238253.396508616
NULL
Cox zph test statistic: 14728.7767838593
NULL
Iteration: 16
NULL
Spec: susinidum_oh *ns(time, 4)
NULL
AIC: 236653.226858842
NULL
Cox zph test statistic: 19308.1808541226
NULL
Iteration: 17
NULL
Spec: susinidum_oh *ns(time, 5)
NULL
AIC: 236251.069626771
NULL
Cox zph test statistic: 20993.031671471
NULL
Iteration: 18
NULL
Spec: susinidum_oh *ns(time, 6)
NULL
AIC: 235896.939555169
NULL
Cox zph test statistic: 21101.9090902985
NULL
Iteration: 19
NULL
Spec: susinidum_oh *rcs(time, 3)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULL
X matrix deemed to be singular; variable time''
Iteration: 20
NULL
Spec: susinidum_oh *rcs(time, 4)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULL
X matrix deemed to be singular; variable time'' time'''
Iteration: 21
NULL
Spec: susinidum_oh *rcs(time, 5)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULL
X matrix deemed to be singular; variable time'' time'''
Iteration: 22
NULL
Spec: susinidum_oh *rcs(time, 6)
NULL
AIC: 239671.217452014
NULL
Cox zph test statistic: 9313.41322424369
NULLCode
results_susini_oh2[which(as.numeric(results_susini_oh2$cox_zph)==
min(as.numeric(results_susini_oh2$cox_zph))),] spec aic cox_zph success
2 susinidum_oh *log(time+0.001) 241724.4 1088.605 TRUECode
# spec aic cox_zph success
# 2 susinidum_oh *log(time+0.001) 241712.3 1088.12 TRUE
invisible("It worsen. We do nothing")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_susinidum_oh2_alt <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq(1,5,.1)) {
change_points <- c(5*i, 10*i, 15*i) # Example change points (months)
coefficients <- c(0.5, 1.5, 0.5, 1) # Example coefficients
# Add the time-varying coefficient to your model
data_mine_miss_restr_proc2$susinidum_oh2_rec2 <- with(data_mine_miss_restr_proc2,
susinidum_oh*time_varying_coeff(time, change_points, coefficients))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh2_alt<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh2_rec2 +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh2_alt)
invisible("")
print(paste0("iteration: ",i))
aic_iter <- round(extractAIC(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh2_alt)[[2]],1)
print( paste0("Time points: ",change_points[1],", ", change_points[2], ", ",change_points[3]))
print( paste0("AIC: ", aic_iter))
cox_zph_iter <- format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh2_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_oh2_alt)$table), 1])
print(paste0("Cox ZPH: ", cox_zph_iter))
results_susinidum_oh2_alt <- rbind.data.frame(results_susinidum_oh2_alt,
cbind.data.frame(
spec = paste0("iteration: ",i, "; change points: ", change_points[1],", ", change_points[2], ", ",change_points[3]),
aic = aic_iter,
cox_zph = cox_zph_iter,
success = T
))
}[1] "iteration: 1"
[1] "Time points: 5, 10, 15"
[1] "AIC: 242713.4"
[1] "Cox ZPH: 346.5879"
[1] "iteration: 1.1"
[1] "Time points: 5.5, 11, 16.5"
[1] "AIC: 242719.6"
[1] "Cox ZPH: 326.4351"
[1] "iteration: 1.2"
[1] "Time points: 6, 12, 18"
[1] "AIC: 242716.7"
[1] "Cox ZPH: 322.6523"
[1] "iteration: 1.3"
[1] "Time points: 6.5, 13, 19.5"
[1] "AIC: 242727.5"
[1] "Cox ZPH: 323.4345"
[1] "iteration: 1.4"
[1] "Time points: 7, 14, 21"
[1] "AIC: 242733.8"
[1] "Cox ZPH: 337.7665"
[1] "iteration: 1.5"
[1] "Time points: 7.5, 15, 22.5"
[1] "AIC: 242738.1"
[1] "Cox ZPH: 364.8138"
[1] "iteration: 1.6"
[1] "Time points: 8, 16, 24"
[1] "AIC: 242739.8"
[1] "Cox ZPH: 379.663"
[1] "iteration: 1.7"
[1] "Time points: 8.5, 17, 25.5"
[1] "AIC: 242740.4"
[1] "Cox ZPH: 406.437"
[1] "iteration: 1.8"
[1] "Time points: 9, 18, 27"
[1] "AIC: 242739.5"
[1] "Cox ZPH: 419.1016"
[1] "iteration: 1.9"
[1] "Time points: 9.5, 19, 28.5"
[1] "AIC: 242737.2"
[1] "Cox ZPH: 434.1251"
[1] "iteration: 2"
[1] "Time points: 10, 20, 30"
[1] "AIC: 242733"
[1] "Cox ZPH: 442.814"
[1] "iteration: 2.1"
[1] "Time points: 10.5, 21, 31.5"
[1] "AIC: 242730"
[1] "Cox ZPH: 420.1959"
[1] "iteration: 2.2"
[1] "Time points: 11, 22, 33"
[1] "AIC: 242719.9"
[1] "Cox ZPH: 405.4619"
[1] "iteration: 2.3"
[1] "Time points: 11.5, 23, 34.5"
[1] "AIC: 242715.8"
[1] "Cox ZPH: 382.8406"
[1] "iteration: 2.4"
[1] "Time points: 12, 24, 36"
[1] "AIC: 242713.8"
[1] "Cox ZPH: 361.1335"
[1] "iteration: 2.5"
[1] "Time points: 12.5, 25, 37.5"
[1] "AIC: 242714.2"
[1] "Cox ZPH: 338.2163"
[1] "iteration: 2.6"
[1] "Time points: 13, 26, 39"
[1] "AIC: 242714.9"
[1] "Cox ZPH: 323.4996"
[1] "iteration: 2.7"
[1] "Time points: 13.5, 27, 40.5"
[1] "AIC: 242718.1"
[1] "Cox ZPH: 322.4792"
[1] "iteration: 2.8"
[1] "Time points: 14, 28, 42"
[1] "AIC: 242722.4"
[1] "Cox ZPH: 333.2702"
[1] "iteration: 2.9"
[1] "Time points: 14.5, 29, 43.5"
[1] "AIC: 242727"
[1] "Cox ZPH: 359.1076"
[1] "iteration: 3"
[1] "Time points: 15, 30, 45"
[1] "AIC: 242731.8"
[1] "Cox ZPH: 408.2805"
[1] "iteration: 3.1"
[1] "Time points: 15.5, 31, 46.5"
[1] "AIC: 242735.1"
[1] "Cox ZPH: 475.1458"
[1] "iteration: 3.2"
[1] "Time points: 16, 32, 48"
[1] "AIC: 242739"
[1] "Cox ZPH: 548.163"
[1] "iteration: 3.3"
[1] "Time points: 16.5, 33, 49.5"
[1] "AIC: 242740.2"
[1] "Cox ZPH: 628.4228"
[1] "iteration: 3.4"
[1] "Time points: 17, 34, 51"
[1] "AIC: 242740.2"
[1] "Cox ZPH: 715.4628"
[1] "iteration: 3.5"
[1] "Time points: 17.5, 35, 52.5"
[1] "AIC: 242738.7"
[1] "Cox ZPH: 801.7517"
[1] "iteration: 3.6"
[1] "Time points: 18, 36, 54"
[1] "AIC: 242734.7"
[1] "Cox ZPH: 899.7736"
[1] "iteration: 3.7"
[1] "Time points: 18.5, 37, 55.5"
[1] "AIC: 242731.9"
[1] "Cox ZPH: 982.7124"
[1] "iteration: 3.8"
[1] "Time points: 19, 38, 57"
[1] "AIC: 242723.8"
[1] "Cox ZPH: 1115.342"
[1] "iteration: 3.9"
[1] "Time points: 19.5, 39, 58.5"
[1] "AIC: 242713.1"
[1] "Cox ZPH: 1233.031"
[1] "iteration: 4"
[1] "Time points: 20, 40, 60"
[1] "AIC: 242698.8"
[1] "Cox ZPH: 1364.041"
[1] "iteration: 4.1"
[1] "Time points: 20.5, 41, 61.5"
[1] "AIC: 242688.1"
[1] "Cox ZPH: 1463.879"
[1] "iteration: 4.2"
[1] "Time points: 21, 42, 63"
[1] "AIC: 242672.2"
[1] "Cox ZPH: 1584.966"
[1] "iteration: 4.3"
[1] "Time points: 21.5, 43, 64.5"
[1] "AIC: 242659.5"
[1] "Cox ZPH: 1694.048"
[1] "iteration: 4.4"
[1] "Time points: 22, 44, 66"
[1] "AIC: 242631.3"
[1] "Cox ZPH: 1785.864"
[1] "iteration: 4.5"
[1] "Time points: 22.5, 45, 67.5"
[1] "AIC: 242619.1"
[1] "Cox ZPH: 1860.616"
[1] "iteration: 4.6"
[1] "Time points: 23, 46, 69"
[1] "AIC: 242597.9"
[1] "Cox ZPH: 1920.634"
[1] "iteration: 4.7"
[1] "Time points: 23.5, 47, 70.5"
[1] "AIC: 242583"
[1] "Cox ZPH: 2004.666"
[1] "iteration: 4.8"
[1] "Time points: 24, 48, 72"
[1] "AIC: 242566.1"
[1] "Cox ZPH: 2109.5"
[1] "iteration: 4.9"
[1] "Time points: 24.5, 49, 73.5"
[1] "AIC: 242550.8"
[1] "Cox ZPH: 2180.038"
[1] "iteration: 5"
[1] "Time points: 25, 50, 75"
[1] "AIC: 242534.7"
[1] "Cox ZPH: 2241.884"Code
results_susinidum_oh2_alt[which(as.numeric(results_susinidum_oh2_alt$cox_zph)==
min(as.numeric(results_susinidum_oh2_alt$cox_zph))),] spec aic cox_zph success
18 iteration: 2.7; change points: 13.5, 27, 40.5 242718.1 322.4792 TRUECode
# spec aic cox_zph success
# 18 iteration: 2.7; change points: 13.5, 27, 40.5 242706.4 322.13 TRUE
message(paste0("The minimum of the previous model: ",
format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)$table), 1]),
" vs. the minimum model: ", results_susinidum_oh2_alt[which(as.numeric(results_susinidum_oh2_alt$cox_zph)== min(as.numeric(results_susinidum_oh2_alt$cox_zph))),"cox_zph"]
))Code
#The minimum of the first modified model: 321.3004 vs. the minimum model: 328.8011
invisible("It worsens. We do not integrate this model")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Analyse the violation of proportional hazards")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
try(dev.off())null device
1 Code
z_240422_6_2_restr <- loess(res240422_6_restr[,"psycom_dum_with"] ~ time240422_6_restr, span=0.50) # residuals
par(mfrow= c(3,1))
plot(time240422_6_restr, fitted(z_240422_6_2_restr))
lines(supsmu(time240422_6_restr, res240422_6_restr[,"psycom_dum_with"]),lty=2)
title("Sch. res., psycom_dum_with")
plot(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)[which(attr(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt$coefficients, "names") == "psycom_dum_with")], lwd=2, main= "Residuals of psycom_dum_with")
abline(0,0, col="red", lty=3, lwd=2)
abline(h=model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt$coefficients["psycom_dum_with"], col=3, lwd=2, lty= 3)
legend("bottomleft", legend= c("Reference line for the null effect", "Average hazard over time", "Time-varying hazard"), lty=c(3,2,1), col= c("red",3,1), lwd=2, bty = "n", bg = "transparent")
km_psycom_dum_with <- survfit(Surv(lag_time,time,event)~ psycom_dum_with+
cluster(id), data = data_mine_miss_restr_proc2)
plot(km_psycom_dum_with, fun = "cloglog", xlab = "Time (in days) using log",col=c("black", "red"), ylab = "log-log survival", main = "log-log curves by psycom_dum_with")
legend("bottomright", legend=c("=0", "=1"), col=c("black", "red"), lty=1, bty = "n", bg = "transparent")
recorded_plot <- recordPlot()
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
png(paste0(folder_path,"zph/proportionallity_psycom_dum_with_restr.png"), height=13, width=10, res=500, units="in")
recorded_plot
dev.off()pdf
2 Code
invisible("Now with Initial substance: Alcohol")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_psycom_dum_with <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq(.1,1.5,.025)) {
tvcs <- c(20,80)*i
data_mine_miss_restr_proc2$psycom_dum_with_rec<-
create_time_varying_effect(data_mine_miss_restr_proc2$psycom_dum_with,
data_mine_miss_restr_proc2$time,
tvcs[1],
tvcs[2],
max(data_mine_miss_restr_proc2$time))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with)
invisible("")
print(paste0("iteration: ",i))
aic_iter <- round(extractAIC(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with)[[2]],1)
print( paste0("Time points: ",tvcs[1],", ", tvcs[2]))
print( paste0("AIC: ", aic_iter))
cox_zph_iter <- format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with)$table), 1])
print(paste0("Cox ZPH: ", cox_zph_iter))
results_psycom_dum_with <- rbind.data.frame(results_psycom_dum_with,
cbind.data.frame(
spec = paste0("iteration: ",i, "; tvcs: ", tvcs[1],", ", tvcs[2]),
aic = aic_iter,
cox_zph = cox_zph_iter,
success = T
))
}[1] "iteration: 0.1"
[1] "Time points: 2, 8"
[1] "AIC: 241851.2"
[1] "Cox ZPH: 1883.112"
[1] "iteration: 0.125"
[1] "Time points: 2.5, 10"
[1] "AIC: 241793.5"
[1] "Cox ZPH: 2047.944"
[1] "iteration: 0.15"
[1] "Time points: 3, 12"
[1] "AIC: 241741.1"
[1] "Cox ZPH: 2185.512"
[1] "iteration: 0.175"
[1] "Time points: 3.5, 14"
[1] "AIC: 241695.6"
[1] "Cox ZPH: 2289.825"
[1] "iteration: 0.2"
[1] "Time points: 4, 16"
[1] "AIC: 241659.1"
[1] "Cox ZPH: 2351.071"
[1] "iteration: 0.225"
[1] "Time points: 4.5, 18"
[1] "AIC: 241631.3"
[1] "Cox ZPH: 2370.834"
[1] "iteration: 0.25"
[1] "Time points: 5, 20"
[1] "AIC: 241612.3"
[1] "Cox ZPH: 2350.685"
[1] "iteration: 0.275"
[1] "Time points: 5.5, 22"
[1] "AIC: 241606.1"
[1] "Cox ZPH: 2272.468"
[1] "iteration: 0.3"
[1] "Time points: 6, 24"
[1] "AIC: 241612.6"
[1] "Cox ZPH: 2133.379"
[1] "iteration: 0.325"
[1] "Time points: 6.5, 26"
[1] "AIC: 241629.3"
[1] "Cox ZPH: 1950.946"
[1] "iteration: 0.35"
[1] "Time points: 7, 28"
[1] "AIC: 241658.1"
[1] "Cox ZPH: 1722"
[1] "iteration: 0.375"
[1] "Time points: 7.5, 30"
[1] "AIC: 241699.2"
[1] "Cox ZPH: 1448.429"
[1] "iteration: 0.4"
[1] "Time points: 8, 32"
[1] "AIC: 241749.9"
[1] "Cox ZPH: 1153.775"
[1] "iteration: 0.425"
[1] "Time points: 8.5, 34"
[1] "AIC: 241808.6"
[1] "Cox ZPH: 864.9711"
[1] "iteration: 0.45"
[1] "Time points: 9, 36"
[1] "AIC: 241873.7"
[1] "Cox ZPH: 612.0001"
[1] "iteration: 0.475"
[1] "Time points: 9.5, 38"
[1] "AIC: 241944.7"
[1] "Cox ZPH: 428.4835"
[1] "iteration: 0.5"
[1] "Time points: 10, 40"
[1] "AIC: 242017.1"
[1] "Cox ZPH: 355.5976"
[1] "iteration: 0.525"
[1] "Time points: 10.5, 42"
[1] "AIC: 242092.7"
[1] "Cox ZPH: 414.2425"
[1] "iteration: 0.55"
[1] "Time points: 11, 44"
[1] "AIC: 242169.9"
[1] "Cox ZPH: 617.0852"
[1] "iteration: 0.575"
[1] "Time points: 11.5, 46"
[1] "AIC: 242247.3"
[1] "Cox ZPH: 957.8698"
[1] "iteration: 0.6"
[1] "Time points: 12, 48"
[1] "AIC: 242323"
[1] "Cox ZPH: 1409.342"
[1] "iteration: 0.625"
[1] "Time points: 12.5, 50"
[1] "AIC: 242398.4"
[1] "Cox ZPH: 1950.015"
[1] "iteration: 0.65"
[1] "Time points: 13, 52"
[1] "AIC: 242471.1"
[1] "Cox ZPH: 2532.404"
[1] "iteration: 0.675"
[1] "Time points: 13.5, 54"
[1] "AIC: 242538.1"
[1] "Cox ZPH: 3100.883"
[1] "iteration: 0.7"
[1] "Time points: 14, 56"
[1] "AIC: 242599"
[1] "Cox ZPH: 3619.677"
[1] "iteration: 0.725"
[1] "Time points: 14.5, 58"
[1] "AIC: 242651.6"
[1] "Cox ZPH: 4060.109"
[1] "iteration: 0.75"
[1] "Time points: 15, 60"
[1] "AIC: 242694.7"
[1] "Cox ZPH: 4405.333"
[1] "iteration: 0.775"
[1] "Time points: 15.5, 62"
[1] "AIC: 242727.3"
[1] "Cox ZPH: 4647.372"
[1] "iteration: 0.8"
[1] "Time points: 16, 64"
[1] "AIC: 242747.8"
[1] "Cox ZPH: 4790.599"
[1] "iteration: 0.825"
[1] "Time points: 16.5, 66"
[1] "AIC: 242754.7"
[1] "Cox ZPH: 4840.749"
[1] "iteration: 0.85"
[1] "Time points: 17, 68"
[1] "AIC: 242746.9"
[1] "Cox ZPH: 4814.885"
[1] "iteration: 0.875"
[1] "Time points: 17.5, 70"
[1] "AIC: 242723.5"
[1] "Cox ZPH: 4725.884"
[1] "iteration: 0.9"
[1] "Time points: 18, 72"
[1] "AIC: 242683.9"
[1] "Cox ZPH: 4587.648"
[1] "iteration: 0.925"
[1] "Time points: 18.5, 74"
[1] "AIC: 242628.3"
[1] "Cox ZPH: 4412.073"
[1] "iteration: 0.95"
[1] "Time points: 19, 76"
[1] "AIC: 242556"
[1] "Cox ZPH: 4211.442"
[1] "iteration: 0.975"
[1] "Time points: 19.5, 78"
[1] "AIC: 242466.1"
[1] "Cox ZPH: 3995.497"
[1] "iteration: 1"
[1] "Time points: 20, 80"
[1] "AIC: 242357.4"
[1] "Cox ZPH: 3766.605"
[1] "iteration: 1.025"
[1] "Time points: 20.5, 82"
[1] "AIC: 242230.8"
[1] "Cox ZPH: 3532.535"
[1] "iteration: 1.05"
[1] "Time points: 21, 84"
[1] "AIC: 242086.1"
[1] "Cox ZPH: 3296.79"
[1] "iteration: 1.075"
[1] "Time points: 21.5, 86"
[1] "AIC: 241928.2"
[1] "Cox ZPH: 3068.773"
[1] "iteration: 1.1"
[1] "Time points: 22, 88"
[1] "AIC: 241757.4"
[1] "Cox ZPH: 2846.697"
[1] "iteration: 1.125"
[1] "Time points: 22.5, 90"
[1] "AIC: 241572.4"
[1] "Cox ZPH: 2629.15"
[1] "iteration: 1.15"
[1] "Time points: 23, 92"
[1] "AIC: 241374.5"
[1] "Cox ZPH: 2418.478"
[1] "iteration: 1.175"
[1] "Time points: 23.5, 94"
[1] "AIC: 241169.8"
[1] "Cox ZPH: 2218.64"
[1] "iteration: 1.2"
[1] "Time points: 24, 96"
[1] "AIC: 240956.5"
[1] "Cox ZPH: 2028.07"
[1] "iteration: 1.225"
[1] "Time points: 24.5, 98"
[1] "AIC: 240741.7"
[1] "Cox ZPH: 1853.869"
[1] "iteration: 1.25"
[1] "Time points: 25, 100"
[1] "AIC: 240521.7"
[1] "Cox ZPH: 1687.786"
[1] "iteration: 1.275"
[1] "Time points: 25.5, 102"
[1] "AIC: 240296.6"
[1] "Cox ZPH: 1534.548"
[1] "iteration: 1.3"
[1] "Time points: 26, 104"
[1] "AIC: 240084.3"
[1] "Cox ZPH: 1405.457"
[1] "iteration: 1.325"
[1] "Time points: 26.5, 106"
[1] "AIC: 239877.9"
[1] "Cox ZPH: 1295.962"
[1] "iteration: 1.35"
[1] "Time points: 27, 108"
[1] "AIC: 239675.8"
[1] "Cox ZPH: 1208.163"
[1] "iteration: 1.375"
[1] "Time points: 27.5, 110"
[1] "AIC: 239479.2"
[1] "Cox ZPH: 1148.109"
[1] "iteration: 1.4"
[1] "Time points: 28, 112"
[1] "AIC: 239290.5"
[1] "Cox ZPH: 1123.353"
[1] "iteration: 1.425"
[1] "Time points: 28.5, 114"
[1] "AIC: 239120.8"
[1] "Cox ZPH: 1142.212"
[1] "iteration: 1.45"
[1] "Time points: 29, 116"
[1] "AIC: 238980.7"
[1] "Cox ZPH: 1210.523"
[1] "iteration: 1.475"
[1] "Time points: 29.5, 118"
[1] "AIC: 238863.6"
[1] "Cox ZPH: 1343.285"
[1] "iteration: 1.5"
[1] "Time points: 30, 120"
[1] "AIC: 238771.2"
[1] "Cox ZPH: 1582.226"Code
results_psycom_dum_with[which(as.numeric(results_psycom_dum_with$cox_zph)==
min(as.numeric(results_psycom_dum_with$cox_zph))),] spec aic cox_zph success
17 iteration: 0.5; tvcs: 10, 40 242017.1 355.5976 TRUECode
# spec aic cox_zph success
# 17 iteration: 0.5; tvcs: 10, 40 242005.3 355.3837 TRUE
message(paste0("The previously modified model: ",
round(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)$table), 1],1),
" vs. the minimum model: ", results_psycom_dum_with[which(as.numeric(results_psycom_dum_with$cox_zph)== min(as.numeric(results_psycom_dum_with$cox_zph))),"cox_zph"]
))Code
#The previously modified model: 317.4 vs. the minimum model: 355.3837
invisible("It worsen")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
require(splines)
specs_psycom_dum_with_rec <- paste0("psycom_dum_with *", c("time", "log(time+0.001)", "poly(ifelse(is.na(time), .001,time), 3)", "poly(ifelse(is.na(time), .001,time), 4)", "poly(ifelse(is.na(time), .001,time), 5)", "poly(ifelse(is.na(time), .001,time), 6)", "pspline(log(time+0.001), df=3)", "pspline(log(time+0.001), df=4)", "pspline(log(time+0.001), df=5)", "pspline(log(time+0.001), df=6)", "bs(time, 3)", "bs(time, 4)", "bs(time, 5)", "bs(time, 6)", "ns(time, 3)", "ns(time, 4)", "ns(time, 5)", "ns(time, 6)", "rcs(time, 3)", "rcs(time, 4)", "rcs(time, 5)", "rcs(time, 6)"))
results_psycom_dum_with_rec <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq_along(specs_psycom_dum_with_rec)) {
l <- specs_psycom_dum_with_rec[i]
result <- tryCatch({
model <- cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
I(round(",l,",2)) +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
dplyr::mutate(log_lag_dias_treat_imp_sin_na_rec = bs(log_lag_dias_treat_imp_sin_na, 3)) %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
aic_val <- extractAIC(model)[[2]]
cox_zph_val <- cox.zph(model)$table[nrow(cox.zph(model)$table), 1]
success <- TRUE
}, error = function(e) {
aic_val <- NA
cox_zph_val <- NA
success <- FALSE
warning(paste0("Error fitting specification ", i, " (", l, "): ", e))
})
results_psycom_dum_with_rec <- rbind(results_psycom_dum_with_rec, data.frame(spec = l, aic = aic_val, cox_zph = cox_zph_val, success = success))
print(cat(paste("Iteration:", i, "\n"))) # Add iteration number
print(cat(paste("Spec: ", l, "\n")))
print(cat(paste("AIC: ", results_psycom_dum_with_rec$aic[i], "\n")))
print(cat(paste("Cox zph test statistic: ", results_psycom_dum_with_rec$cox_zph[i], "\n")))
}Iteration: 1
NULL
Spec: psycom_dum_with *time
NULL
AIC: 242073.646015772
NULL
Cox zph test statistic: 1270.82638037256
NULL
Iteration: 2
NULL
Spec: psycom_dum_with *log(time+0.001)
NULL
AIC: 242679.202046089
NULL
Cox zph test statistic: 495.282058112557
NULL
Iteration: 3
NULL
Spec: psycom_dum_with *poly(ifelse(is.na(time), .001,time), 3)
NULL
AIC: 241653.345412216
NULL
Cox zph test statistic: 5532.18767068606
NULL
Iteration: 4
NULL
Spec: psycom_dum_with *poly(ifelse(is.na(time), .001,time), 4)
NULL
AIC: 241374.872334666
NULL
Cox zph test statistic: 5845.6906904199
NULL
Iteration: 5
NULL
Spec: psycom_dum_with *poly(ifelse(is.na(time), .001,time), 5)
NULL
AIC: 241301.737606574
NULL
Cox zph test statistic: 5895.95894768982
NULL
Iteration: 6
NULL
Spec: psycom_dum_with *poly(ifelse(is.na(time), .001,time), 6)
NULL
AIC: 241133.859850746
NULL
Cox zph test statistic: 5814.08088848928
NULL
Iteration: 7
NULL
Spec: psycom_dum_with *pspline(log(time+0.001), df=3)
NULL
AIC: 241133.859850746
NULL
Cox zph test statistic: 5814.08088848928
NULL
Iteration: 8
NULL
Spec: psycom_dum_with *pspline(log(time+0.001), df=4)
NULL
AIC: 241133.859850746
NULL
Cox zph test statistic: 5814.08088848928
NULL
Iteration: 9
NULL
Spec: psycom_dum_with *pspline(log(time+0.001), df=5)
NULL
AIC: 241133.859850746
NULL
Cox zph test statistic: 5814.08088848928
NULL
Iteration: 10
NULL
Spec: psycom_dum_with *pspline(log(time+0.001), df=6)
NULL
AIC: 241133.859850746
NULL
Cox zph test statistic: 5814.08088848928
NULL
Iteration: 11
NULL
Spec: psycom_dum_with *bs(time, 3)
NULL
AIC: 240806.877886123
NULL
Cox zph test statistic: 8350.2319648494
NULL
Iteration: 12
NULL
Spec: psycom_dum_with *bs(time, 4)
NULL
AIC: 237979.785769352
NULL
Cox zph test statistic: 13977.8757470223
NULL
Iteration: 13
NULL
Spec: psycom_dum_with *bs(time, 5)
NULL
AIC: 237406.174681623
NULL
Cox zph test statistic: 15528.301431604
NULL
Iteration: 14
NULL
Spec: psycom_dum_with *bs(time, 6)
NULL
AIC: 237229.627963973
NULL
Cox zph test statistic: 16040.7359383214
NULL
Iteration: 15
NULL
Spec: psycom_dum_with *ns(time, 3)
NULL
AIC: 239286.89566843
NULL
Cox zph test statistic: 11649.939360018
NULL
Iteration: 16
NULL
Spec: psycom_dum_with *ns(time, 4)
NULL
AIC: 237893.753000825
NULL
Cox zph test statistic: 14575.1889880412
NULL
Iteration: 17
NULL
Spec: psycom_dum_with *ns(time, 5)
NULL
AIC: 237600.950385534
NULL
Cox zph test statistic: 15639.9138673865
NULL
Iteration: 18
NULL
Spec: psycom_dum_with *ns(time, 6)
NULL
AIC: 237354.154406415
NULL
Cox zph test statistic: 15735.8237332929
NULL
Iteration: 19
NULL
Spec: psycom_dum_with *rcs(time, 3)
NULL
AIC: 240998.015320773
NULL
Cox zph test statistic: 7177.11023618446
NULL
X matrix deemed to be singular; variable time''
Iteration: 20
NULL
Spec: psycom_dum_with *rcs(time, 4)
NULL
AIC: 240998.015320773
NULL
Cox zph test statistic: 7177.11023618446
NULL
X matrix deemed to be singular; variable time'' time'''
Iteration: 21
NULL
Spec: psycom_dum_with *rcs(time, 5)
NULL
AIC: 240998.015320773
NULL
Cox zph test statistic: 7177.11023618446
NULL
X matrix deemed to be singular; variable time'' time'''
Iteration: 22
NULL
Spec: psycom_dum_with *rcs(time, 6)
NULL
AIC: 240998.015320773
NULL
Cox zph test statistic: 7177.11023618446
NULLCode
results_psycom_dum_with_rec[which(as.numeric(results_psycom_dum_with_rec$cox_zph)==
min(as.numeric(results_psycom_dum_with_rec$cox_zph))),] spec aic cox_zph success
2 psycom_dum_with *log(time+0.001) 242679.2 495.2821 TRUECode
# spec aic cox_zph success
# 2 psycom_dum_with *log(time+0.001) 242667.5 494.8698 TRUE
invisible("It worsen. We do nothing")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_psycom_dum_with_alt <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
for (i in seq(1,5,.1)) {
change_points <- c(5*i, 10*i, 15*i) # Example change points (months)
coefficients <- c(0.5, 1.5, 0.5, 1) # Example coefficients
# Add the time-varying coefficient to your model
data_mine_miss_restr_proc2$psycom_dum_with_rec2 <- with(data_mine_miss_restr_proc2,
psycom_dum_with*time_varying_coeff(time, change_points, coefficients))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with_alt<-
cph(as.formula(paste0("Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod)")),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with_alt)
invisible("")
print(paste0("iteration: ",i))
aic_iter <- round(extractAIC(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with_alt)[[2]],1)
print( paste0("Time points: ",change_points[1],", ", change_points[2], ", ",change_points[3]))
print( paste0("AIC: ", aic_iter))
cox_zph_iter <- format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with_alt)$table), 1])
print(paste0("Cox ZPH: ", cox_zph_iter))
results_psycom_dum_with_alt <- rbind.data.frame(results_psycom_dum_with_alt,
cbind.data.frame(
spec = paste0("iteration: ",i, "; change points: ", change_points[1],", ", change_points[2], ", ",change_points[3]),
aic = aic_iter,
cox_zph = cox_zph_iter,
success = T
))
}[1] "iteration: 1"
[1] "Time points: 5, 10, 15"
[1] "AIC: 242751.2"
[1] "Cox ZPH: 328.9806"
[1] "iteration: 1.1"
[1] "Time points: 5.5, 11, 16.5"
[1] "AIC: 242750.6"
[1] "Cox ZPH: 320.86"
[1] "iteration: 1.2"
[1] "Time points: 6, 12, 18"
[1] "AIC: 242750.2"
[1] "Cox ZPH: 315.1147"
[1] "iteration: 1.3"
[1] "Time points: 6.5, 13, 19.5"
[1] "AIC: 242747.3"
[1] "Cox ZPH: 308.1648"
[1] "iteration: 1.4"
[1] "Time points: 7, 14, 21"
[1] "AIC: 242746.4"
[1] "Cox ZPH: 306.6321"
[1] "iteration: 1.5"
[1] "Time points: 7.5, 15, 22.5"
[1] "AIC: 242745.3"
[1] "Cox ZPH: 306.8176"
[1] "iteration: 1.6"
[1] "Time points: 8, 16, 24"
[1] "AIC: 242742.7"
[1] "Cox ZPH: 307.9781"
[1] "iteration: 1.7"
[1] "Time points: 8.5, 17, 25.5"
[1] "AIC: 242739.7"
[1] "Cox ZPH: 310.7349"
[1] "iteration: 1.8"
[1] "Time points: 9, 18, 27"
[1] "AIC: 242740.6"
[1] "Cox ZPH: 311.398"
[1] "iteration: 1.9"
[1] "Time points: 9.5, 19, 28.5"
[1] "AIC: 242738.9"
[1] "Cox ZPH: 312.5175"
[1] "iteration: 2"
[1] "Time points: 10, 20, 30"
[1] "AIC: 242732.5"
[1] "Cox ZPH: 314.7717"
[1] "iteration: 2.1"
[1] "Time points: 10.5, 21, 31.5"
[1] "AIC: 242731.2"
[1] "Cox ZPH: 313.7358"
[1] "iteration: 2.2"
[1] "Time points: 11, 22, 33"
[1] "AIC: 242725.9"
[1] "Cox ZPH: 314.3701"
[1] "iteration: 2.3"
[1] "Time points: 11.5, 23, 34.5"
[1] "AIC: 242724.4"
[1] "Cox ZPH: 312.8812"
[1] "iteration: 2.4"
[1] "Time points: 12, 24, 36"
[1] "AIC: 242723"
[1] "Cox ZPH: 311.2805"
[1] "iteration: 2.5"
[1] "Time points: 12.5, 25, 37.5"
[1] "AIC: 242724.9"
[1] "Cox ZPH: 310.5611"
[1] "iteration: 2.6"
[1] "Time points: 13, 26, 39"
[1] "AIC: 242720.3"
[1] "Cox ZPH: 312.4336"
[1] "iteration: 2.7"
[1] "Time points: 13.5, 27, 40.5"
[1] "AIC: 242719.7"
[1] "Cox ZPH: 316.791"
[1] "iteration: 2.8"
[1] "Time points: 14, 28, 42"
[1] "AIC: 242719.8"
[1] "Cox ZPH: 325.1106"
[1] "iteration: 2.9"
[1] "Time points: 14.5, 29, 43.5"
[1] "AIC: 242724.8"
[1] "Cox ZPH: 336.6359"
[1] "iteration: 3"
[1] "Time points: 15, 30, 45"
[1] "AIC: 242727.3"
[1] "Cox ZPH: 351.6382"
[1] "iteration: 3.1"
[1] "Time points: 15.5, 31, 46.5"
[1] "AIC: 242730.9"
[1] "Cox ZPH: 372.395"
[1] "iteration: 3.2"
[1] "Time points: 16, 32, 48"
[1] "AIC: 242736.4"
[1] "Cox ZPH: 399.6994"
[1] "iteration: 3.3"
[1] "Time points: 16.5, 33, 49.5"
[1] "AIC: 242739.6"
[1] "Cox ZPH: 428.2439"
[1] "iteration: 3.4"
[1] "Time points: 17, 34, 51"
[1] "AIC: 242744.2"
[1] "Cox ZPH: 468.6018"
[1] "iteration: 3.5"
[1] "Time points: 17.5, 35, 52.5"
[1] "AIC: 242746.8"
[1] "Cox ZPH: 502.0081"
[1] "iteration: 3.6"
[1] "Time points: 18, 36, 54"
[1] "AIC: 242750"
[1] "Cox ZPH: 538.8976"
[1] "iteration: 3.7"
[1] "Time points: 18.5, 37, 55.5"
[1] "AIC: 242751.2"
[1] "Cox ZPH: 571.5965"
[1] "iteration: 3.8"
[1] "Time points: 19, 38, 57"
[1] "AIC: 242753.3"
[1] "Cox ZPH: 613.0588"
[1] "iteration: 3.9"
[1] "Time points: 19.5, 39, 58.5"
[1] "AIC: 242754.5"
[1] "Cox ZPH: 653.7076"
[1] "iteration: 4"
[1] "Time points: 20, 40, 60"
[1] "AIC: 242754.7"
[1] "Cox ZPH: 711.0753"
[1] "iteration: 4.1"
[1] "Time points: 20.5, 41, 61.5"
[1] "AIC: 242754.3"
[1] "Cox ZPH: 751.489"
[1] "iteration: 4.2"
[1] "Time points: 21, 42, 63"
[1] "AIC: 242752.7"
[1] "Cox ZPH: 801.2111"
[1] "iteration: 4.3"
[1] "Time points: 21.5, 43, 64.5"
[1] "AIC: 242750.3"
[1] "Cox ZPH: 852.915"
[1] "iteration: 4.4"
[1] "Time points: 22, 44, 66"
[1] "AIC: 242746.5"
[1] "Cox ZPH: 899.0964"
[1] "iteration: 4.5"
[1] "Time points: 22.5, 45, 67.5"
[1] "AIC: 242743.9"
[1] "Cox ZPH: 938.8631"
[1] "iteration: 4.6"
[1] "Time points: 23, 46, 69"
[1] "AIC: 242741.7"
[1] "Cox ZPH: 961.7692"
[1] "iteration: 4.7"
[1] "Time points: 23.5, 47, 70.5"
[1] "AIC: 242738.3"
[1] "Cox ZPH: 1000.855"
[1] "iteration: 4.8"
[1] "Time points: 24, 48, 72"
[1] "AIC: 242733.8"
[1] "Cox ZPH: 1045.778"
[1] "iteration: 4.9"
[1] "Time points: 24.5, 49, 73.5"
[1] "AIC: 242729.4"
[1] "Cox ZPH: 1075.742"
[1] "iteration: 5"
[1] "Time points: 25, 50, 75"
[1] "AIC: 242725.9"
[1] "Cox ZPH: 1113.625"Code
results_psycom_dum_with_alt[which(as.numeric(results_psycom_dum_with_alt$cox_zph)==
min(as.numeric(results_psycom_dum_with_alt$cox_zph))),] spec aic cox_zph success
5 iteration: 1.4; change points: 7, 14, 21 242746.4 306.6321 TRUECode
# spec aic cox_zph success
# 5 iteration: 1.4; change points: 7, 14, 21 242734.7 306.2739 TRUE
message(paste0("The minimum of the previous model: ",
format(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)$table[nrow(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt)$table), 1]),
" vs. the minimum model: ", results_psycom_dum_with_alt[which(as.numeric(results_psycom_dum_with_alt$cox_zph)== min(as.numeric(results_psycom_dum_with_alt$cox_zph))),"cox_zph"]
))Code
#The minimum of the previous model: 317.3692 vs. the minimum model: 306.2739
invisible("It worsens. We do not integrate this model")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
results_psycom_dum_with_alt[which(as.numeric(results_psycom_dum_with_alt$cox_zph)==
min(as.numeric(results_psycom_dum_with_alt$cox_zph))),] spec aic cox_zph success
5 iteration: 1.4; change points: 7, 14, 21 242746.4 306.6321 TRUECode
change_points <- c(5, 10, 15)*1.4 # Example change points (months)
coefficients <- c(0.5, 1.5, 0.5, 1) # Example coefficients
# Add the time-varying coefficient to your model
data_mine_miss_restr_proc2$psycom_dum_with_rec2 <- with(data_mine_miss_restr_proc2,
psycom_dum_with*time_varying_coeff(time, change_points, coefficients))
model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with_alt<-
cph(Surv(lag_time,time,event==1)~
cluster(id)+
lag_tr_outcome_rec +
log_lag_dias_treat_imp_sin_na +
lag_less_90d_tr1_rec+
lag_comp_bpsc_y3_severe_rec +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
strat(tipo_de_plan_2_mod),
data=data_mine_miss_restr_proc2 %>%
data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
round(cox.zph(model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with_alt)$table,4) chisq df p
lag_tr_outcome_rec 66.0738 1 0.0000
log_lag_dias_treat_imp_sin_na 10.3740 1 0.0013
lag_less_90d_tr1_rec 83.7240 1 0.0000
lag_comp_bpsc_y3_severe_rec 17.2145 1 0.0000
lag_policonsumo2 3.2530 1 0.0713
edad_al_ing_1 0.1557 1 0.6931
ano_nac_corr 5.1059 1 0.0238
susinidum_coc_rec2 2.9522 1 0.0858
susinidum_oh 19.8355 1 0.0000
susinidum_pbc 2.0732 1 0.1499
susinidum_mar 6.7703 1 0.0093
psycom_dum_with_rec2 0.7582 1 0.3839
psycom_dum_study 1.4761 1 0.2244
freq_cons_dum_5day 0.0802 1 0.7770
cond_oc_dum_2inact 4.4838 1 0.0342
cond_oc_dum_3unemp 1.7238 1 0.1892
susprindum_oh 4.4440 1 0.0350
susprindum_coc 1.5636 1 0.2111
susprindum_pbc 0.0418 1 0.8380
susprindum_mar 5.2227 1 0.0223
GLOBAL 306.6321 20 0.0000Code
# chisq df p
# lag_tr_outcome_rec 66.1801 1 0.0000 ####
# log_lag_dias_treat_imp_sin_na 10.2171 1 0.0014 #
# lag_less_90d_tr1_rec 83.7258 1 0.0000 ####
# lag_comp_bpsc_y3_severe_rec 17.1686 1 0.0000 ####
# lag_policonsumo2 3.2220 1 0.0727
# edad_al_ing_1 0.1476 1 0.7008
# ano_nac_corr 5.1430 1 0.0233
# susinidum_coc_rec2 2.9413 1 0.0863
# susinidum_oh 19.6325 1 0.0000 ####
# susinidum_pbc 1.9940 1 0.1579
# susinidum_mar 6.7367 1 0.0094 #
# psycom_dum_with_rec2 0.7771 1 0.3780
# psycom_dum_study 1.4873 1 0.2226
# freq_cons_dum_5day 0.0522 1 0.8193
# cond_oc_dum_2inact 4.5082 1 0.0337
# cond_oc_dum_3unemp 1.6053 1 0.2051
# susprindum_oh 4.3562 1 0.0369
# susprindum_coc 1.6009 1 0.2058
# susprindum_pbc 0.0568 1 0.8117
# susprindum_mar 5.2723 1 0.0217
# GLOBAL 306.2739 20 0.0000
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Analyse the violation of proportional hazards")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_Apply
Code
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
invisible("IIW")
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
levels_tipo_de_plan_2_mod <-attr(table(data_mine_miss_restr_proc2$tipo_de_plan_2_mod),"name")
invisible("model_restr_after_mod_not_log_lag_dias_treat_lag_less_90d_bpsc_severe_alt_tr_out_susinidum_coc_alt_psycom_dum_with_alt, make IIWs")
invisible("basic ambulatory")
iiw_model_after_ph_ba<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1_rec.lag+
comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod== levels_tipo_de_plan_2_mod[[1]]) %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= subset(maxfu_restr_df, subset= maxfu.id %in% subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod==levels_tipo_de_plan_2_mod[[1]])$id) %>% as.data.frame(),
# formulanull = Surv(time.lag,time,event)~ edad_al_ing_1 +
# ano_nac_corr,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc_rec2", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with_rec2", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome_rec", "comp_bpsc_y3_severe_rec", "less_90d_tr1_rec", "policonsumo2"),
lagfirst= c(2.95082,4.499811/2,0,0,0,0), #90/30.5 4.499811 es 90 días
first= T
)
invisible("2024-05-02: problem with different values in people with plans (possibly imputed)")
dis_run_w_more_type_plan<-
data_mine_miss_restr_proc2 %>%
dplyr::group_by(id) %>%
dplyr::summarise(count_n_dis=n_distinct(tipo_de_plan_2_mod)) %>%
dplyr::filter(count_n_dis>1) %>% distinct(id)
summary(iiw_model_after_ph_ba$m)Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 13471, number of events= 9113
(880 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome_rec.lag 0.170110 1.185436 0.055701 0.082986 2.050
log_dias_treat_imp_sin_na.lag -0.236237 0.789594 0.006277 0.022273 -10.606
less_90d_tr1_rec.lag -0.112838 0.893295 0.070938 0.086272 -1.308
comp_bpsc_y3_severe_rec.lag 0.034013 1.034598 0.092579 0.074490 0.457
policonsumo2.lag -0.636014 0.529398 0.025642 0.026439 -24.056
edad_al_ing_1 0.039205 1.039984 0.005187 0.003358 11.675
ano_nac_corr 0.044017 1.045000 0.005112 0.003333 13.207
susinidum_coc_rec2 0.166251 1.180869 0.082712 0.064699 2.570
susinidum_oh 0.102885 1.108363 0.064329 0.049501 2.078
susinidum_pbc -0.016869 0.983272 0.076770 0.057891 -0.291
susinidum_mar 0.097263 1.102151 0.065595 0.052648 1.847
psycom_dum_with_rec2 0.015614 1.015737 0.021836 0.014045 1.112
psycom_dum_study -0.182551 0.833142 0.032157 0.030477 -5.990
freq_cons_dum_5day -0.016709 0.983430 0.024123 0.020375 -0.820
cond_oc_dum_2inact -0.002615 0.997388 0.033110 0.025535 -0.102
cond_oc_dum_3unemp 0.066169 1.068408 0.025464 0.017982 3.680
susprindum_oh -0.019388 0.980799 0.097556 0.059609 -0.325
susprindum_coc 0.101071 1.106356 0.098804 0.060316 1.676
susprindum_pbc 0.118128 1.125388 0.097381 0.060901 1.940
susprindum_mar 0.122394 1.130199 0.103347 0.062593 1.955
Pr(>|z|)
tr_outcome_rec.lag 0.040377 *
log_dias_treat_imp_sin_na.lag < 2e-16 ***
less_90d_tr1_rec.lag 0.190897
comp_bpsc_y3_severe_rec.lag 0.647949
policonsumo2.lag < 2e-16 ***
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_coc_rec2 0.010181 *
susinidum_oh 0.037671 *
susinidum_pbc 0.770749
susinidum_mar 0.064684 .
psycom_dum_with_rec2 0.266259
psycom_dum_study 2.1e-09 ***
freq_cons_dum_5day 0.412174
cond_oc_dum_2inact 0.918429
cond_oc_dum_3unemp 0.000233 ***
susprindum_oh 0.744993
susprindum_coc 0.093799 .
susprindum_pbc 0.052420 .
susprindum_mar 0.050537 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome_rec.lag 1.1854 0.8436 1.0075 1.3948
log_dias_treat_imp_sin_na.lag 0.7896 1.2665 0.7559 0.8248
less_90d_tr1_rec.lag 0.8933 1.1195 0.7543 1.0579
comp_bpsc_y3_severe_rec.lag 1.0346 0.9666 0.8941 1.1972
policonsumo2.lag 0.5294 1.8889 0.5027 0.5576
edad_al_ing_1 1.0400 0.9616 1.0332 1.0469
ano_nac_corr 1.0450 0.9569 1.0382 1.0518
susinidum_coc_rec2 1.1809 0.8468 1.0402 1.3405
susinidum_oh 1.1084 0.9022 1.0059 1.2213
susinidum_pbc 0.9833 1.0170 0.8778 1.1014
susinidum_mar 1.1022 0.9073 0.9941 1.2220
psycom_dum_with_rec2 1.0157 0.9845 0.9882 1.0441
psycom_dum_study 0.8331 1.2003 0.7848 0.8844
freq_cons_dum_5day 0.9834 1.0168 0.9449 1.0235
cond_oc_dum_2inact 0.9974 1.0026 0.9487 1.0486
cond_oc_dum_3unemp 1.0684 0.9360 1.0314 1.1067
susprindum_oh 0.9808 1.0196 0.8727 1.1023
susprindum_coc 1.1064 0.9039 0.9830 1.2452
susprindum_pbc 1.1254 0.8886 0.9988 1.2681
susprindum_mar 1.1302 0.8848 0.9997 1.2777
Concordance= 0.635 (se = 0.002 )
Likelihood ratio test= 2204 on 20 df, p=<2e-16
Wald test = 1947 on 20 df, p=<2e-16
Score (logrank) test = 2554 on 20 df, p=<2e-16, Robust = 2679 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).Code
# Call:
# coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
# log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
# policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
# susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar, data = datacox, cluster = id)
#
# n= 13471, number of events= 9113
# (880 observations deleted due to missingness)
#
# coef exp(coef) se(coef) robust se z Pr(>|z|)
# tr_outcome_rec.lag -0.053168 0.948220 0.057677 0.039772 -1.337 0.18128
# log_dias_treat_imp_sin_na.lag -0.038928 0.961820 0.013773 0.015230 -2.556 0.01059 *
# less_90d_tr1_rec.lag 0.095160 1.099835 0.073449 0.053940 1.764 0.07770 .
# comp_bpsc_y3_severe_rec.lag 0.009360 1.009404 0.092595 0.069811 0.134 0.89335
# policonsumo2.lag -0.799276 0.449654 0.024645 0.019765 -40.438 < 2e-16 ***
# edad_al_ing_1 0.023181 1.023452 0.005167 0.003189 7.268 3.64e-13 ***
# ano_nac_corr 0.030272 1.030735 0.005099 0.003160 9.581 < 2e-16 ***
# susinidum_coc_rec2 0.113215 1.119873 0.081735 0.057953 1.954 0.05075 .
# susinidum_oh 0.064995 1.067154 0.064022 0.041330 1.573 0.11581
# susinidum_pbc -0.091810 0.912278 0.076500 0.051320 -1.789 0.07362 .
# susinidum_mar 0.070750 1.073313 0.065353 0.042346 1.671 0.09477 .
# psycom_dum_with_rec2 0.009316 1.009359 0.021927 0.012820 0.727 0.46744
# psycom_dum_study -0.172542 0.841523 0.032163 0.024628 -7.006 2.45e-12 ***
# freq_cons_dum_5day 0.002757 1.002760 0.024133 0.014661 0.188 0.85086
# cond_oc_dum_2inact -0.019184 0.980999 0.033110 0.020755 -0.924 0.35532
# cond_oc_dum_3unemp 0.053533 1.054991 0.025441 0.016401 3.264 0.00110 **
# susprindum_oh 0.004280 1.004289 0.097915 0.059290 0.072 0.94245
# susprindum_coc 0.167440 1.182274 0.099029 0.060961 2.747 0.00602 **
# susprindum_pbc 0.166351 1.180988 0.097732 0.060049 2.770 0.00560 **
# susprindum_mar 0.152467 1.164705 0.103688 0.062710 2.431 0.01504 *
# ---
# Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#
# exp(coef) exp(-coef) lower .95 upper .95
# tr_outcome_rec.lag 0.9482 1.0546 0.8771 1.0251
# log_dias_treat_imp_sin_na.lag 0.9618 1.0397 0.9335 0.9910
# less_90d_tr1_rec.lag 1.0998 0.9092 0.9895 1.2225
# comp_bpsc_y3_severe_rec.lag 1.0094 0.9907 0.8803 1.1574
# policonsumo2.lag 0.4497 2.2239 0.4326 0.4674
# edad_al_ing_1 1.0235 0.9771 1.0171 1.0299
# ano_nac_corr 1.0307 0.9702 1.0244 1.0371
# susinidum_coc_rec2 1.1199 0.8930 0.9996 1.2546
# susinidum_oh 1.0672 0.9371 0.9841 1.1572
# susinidum_pbc 0.9123 1.0962 0.8250 1.0088
# susinidum_mar 1.0733 0.9317 0.9878 1.1662
# psycom_dum_with_rec2 1.0094 0.9907 0.9843 1.0350
# psycom_dum_study 0.8415 1.1883 0.8019 0.8831
# freq_cons_dum_5day 1.0028 0.9972 0.9744 1.0320
# cond_oc_dum_2inact 0.9810 1.0194 0.9419 1.0217
# cond_oc_dum_3unemp 1.0550 0.9479 1.0216 1.0895
# susprindum_oh 1.0043 0.9957 0.8941 1.1280
# susprindum_coc 1.1823 0.8458 1.0491 1.3323
# susprindum_pbc 1.1810 0.8467 1.0499 1.3285
# susprindum_mar 1.1647 0.8586 1.0300 1.3170
#
# Concordance= 0.609 (se = 0.002 )
# Likelihood ratio test= 1284 on 20 df, p=<2e-16
# Wald test = 2579 on 20 df, p=<2e-16
# Score (logrank) test = 1333 on 20 df, p=<2e-16, Robust = 2025 p=<2e-16
#
# (Note: the likelihood ratio and score tests assume independence of
# observations within a cluster, the Wald and robust score tests do not).
summary(iiw_model_after_ph_ba$iiw) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.09172 0.55656 1.00031 1.24301 1.82741 5.01220 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.5685 0.8381 1.0387 1.3327 1.8348 3.4639
iiw_model_after_ph_ba_alt<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1_rec.lag+
comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod== levels_tipo_de_plan_2_mod[[1]]) %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= subset(maxfu_restr_df, subset= maxfu.id %in% subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod==levels_tipo_de_plan_2_mod[[1]])$id) %>% as.data.frame(),
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc_rec2", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with_rec2", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome_rec", "comp_bpsc_y3_severe_rec", "less_90d_tr1_rec", "policonsumo2"),
lagfirst= c(2.95082,4.499811,1,1,1,1), #90/30.5 4.499811 es 90 días
first= T
)
summary(iiw_model_after_ph_ba_alt$m)Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 13471, number of events= 9113
(880 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome_rec.lag 2.114216 8.283090 0.072640 0.068485 30.871
log_dias_treat_imp_sin_na.lag 0.484348 1.623117 0.023209 0.028675 16.891
less_90d_tr1_rec.lag 0.399869 1.491630 0.070287 0.062216 6.427
comp_bpsc_y3_severe_rec.lag 0.692117 1.997940 0.055532 0.054833 12.622
policonsumo2.lag -0.009252 0.990791 0.032970 0.027511 -0.336
edad_al_ing_1 0.095935 1.100687 0.005243 0.004801 19.983
ano_nac_corr 0.101247 1.106550 0.005183 0.004728 21.415
susinidum_coc_rec2 0.162364 1.176289 0.080265 0.090627 1.792
susinidum_oh 0.049976 1.051246 0.063197 0.057214 0.873
susinidum_pbc 0.067771 1.070120 0.075944 0.068309 0.992
susinidum_mar 0.080871 1.084231 0.064602 0.058516 1.382
psycom_dum_with_rec2 -0.077395 0.925524 0.021792 0.017770 -4.355
psycom_dum_study -0.171662 0.842263 0.032046 0.030499 -5.629
freq_cons_dum_5day -0.013027 0.987058 0.024045 0.021068 -0.618
cond_oc_dum_2inact -0.005455 0.994559 0.033066 0.028766 -0.190
cond_oc_dum_3unemp 0.047916 1.049082 0.025405 0.022455 2.134
susprindum_oh 0.124032 1.132052 0.097439 0.078021 1.590
susprindum_coc 0.176572 1.193120 0.098636 0.079742 2.214
susprindum_pbc 0.217877 1.243434 0.097414 0.078704 2.768
susprindum_mar 0.183830 1.201811 0.103376 0.083412 2.204
Pr(>|z|)
tr_outcome_rec.lag < 2e-16 ***
log_dias_treat_imp_sin_na.lag < 2e-16 ***
less_90d_tr1_rec.lag 1.30e-10 ***
comp_bpsc_y3_severe_rec.lag < 2e-16 ***
policonsumo2.lag 0.73665
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_coc_rec2 0.07320 .
susinidum_oh 0.38240
susinidum_pbc 0.32114
susinidum_mar 0.16696
psycom_dum_with_rec2 1.33e-05 ***
psycom_dum_study 1.82e-08 ***
freq_cons_dum_5day 0.53636
cond_oc_dum_2inact 0.84959
cond_oc_dum_3unemp 0.03285 *
susprindum_oh 0.11190
susprindum_coc 0.02681 *
susprindum_pbc 0.00563 **
susprindum_mar 0.02753 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome_rec.lag 8.2831 0.1207 7.2427 9.4730
log_dias_treat_imp_sin_na.lag 1.6231 0.6161 1.5344 1.7170
less_90d_tr1_rec.lag 1.4916 0.6704 1.3204 1.6851
comp_bpsc_y3_severe_rec.lag 1.9979 0.5005 1.7944 2.2246
policonsumo2.lag 0.9908 1.0093 0.9388 1.0457
edad_al_ing_1 1.1007 0.9085 1.0904 1.1111
ano_nac_corr 1.1065 0.9037 1.0963 1.1169
susinidum_coc_rec2 1.1763 0.8501 0.9849 1.4049
susinidum_oh 1.0512 0.9513 0.9397 1.1760
susinidum_pbc 1.0701 0.9345 0.9360 1.2234
susinidum_mar 1.0842 0.9223 0.9667 1.2160
psycom_dum_with_rec2 0.9255 1.0805 0.8938 0.9583
psycom_dum_study 0.8423 1.1873 0.7934 0.8941
freq_cons_dum_5day 0.9871 1.0131 0.9471 1.0287
cond_oc_dum_2inact 0.9946 1.0055 0.9400 1.0522
cond_oc_dum_3unemp 1.0491 0.9532 1.0039 1.0963
susprindum_oh 1.1321 0.8834 0.9715 1.3191
susprindum_coc 1.1931 0.8381 1.0205 1.3950
susprindum_pbc 1.2434 0.8042 1.0657 1.4508
susprindum_mar 1.2018 0.8321 1.0206 1.4153
Concordance= 0.706 (se = 0.002 )
Likelihood ratio test= 5371 on 20 df, p=<2e-16
Wald test = 5053 on 20 df, p=<2e-16
Score (logrank) test = 5632 on 20 df, p=<2e-16, Robust = 3864 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).Code
# Call:
# coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
# log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
# policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
# susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar, data = datacox, cluster = id)
#
# n= 13471, number of events= 9113
# (880 observations deleted due to missingness)
#
# coef exp(coef) se(coef) robust se z Pr(>|z|)
# tr_outcome_rec.lag 2.114216 8.283090 0.072640 0.068485 30.871 < 2e-16 ***
# log_dias_treat_imp_sin_na.lag 0.484348 1.623117 0.023209 0.028675 16.891 < 2e-16 ***
# less_90d_tr1_rec.lag 0.399869 1.491630 0.070287 0.062216 6.427 1.30e-10 ***
# comp_bpsc_y3_severe_rec.lag 0.692117 1.997940 0.055532 0.054833 12.622 < 2e-16 ***
# policonsumo2.lag -0.009252 0.990791 0.032970 0.027511 -0.336 0.73665
# edad_al_ing_1 0.095935 1.100687 0.005243 0.004801 19.983 < 2e-16 ***
# ano_nac_corr 0.101247 1.106550 0.005183 0.004728 21.415 < 2e-16 ***
# susinidum_coc_rec2 0.162364 1.176289 0.080265 0.090627 1.792 0.07320 .
# susinidum_oh 0.049976 1.051246 0.063197 0.057214 0.873 0.38240
# susinidum_pbc 0.067771 1.070120 0.075944 0.068309 0.992 0.32114
# susinidum_mar 0.080871 1.084231 0.064602 0.058516 1.382 0.16696
# psycom_dum_with_rec2 -0.077395 0.925524 0.021792 0.017770 -4.355 1.33e-05 ***
# psycom_dum_study -0.171662 0.842263 0.032046 0.030499 -5.629 1.82e-08 ***
# freq_cons_dum_5day -0.013027 0.987058 0.024045 0.021068 -0.618 0.53636
# cond_oc_dum_2inact -0.005455 0.994559 0.033066 0.028766 -0.190 0.84959
# cond_oc_dum_3unemp 0.047916 1.049082 0.025405 0.022455 2.134 0.03285 *
# susprindum_oh 0.124032 1.132052 0.097439 0.078021 1.590 0.11190
# susprindum_coc 0.176572 1.193120 0.098636 0.079742 2.214 0.02681 *
# susprindum_pbc 0.217877 1.243434 0.097414 0.078704 2.768 0.00563 **
# susprindum_mar 0.183830 1.201811 0.103376 0.083412 2.204 0.02753 *
# ---
# Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#
# exp(coef) exp(-coef) lower .95 upper .95
# tr_outcome_rec.lag 8.2831 0.1207 7.2427 9.4730
# log_dias_treat_imp_sin_na.lag 1.6231 0.6161 1.5344 1.7170
# less_90d_tr1_rec.lag 1.4916 0.6704 1.3204 1.6851
# comp_bpsc_y3_severe_rec.lag 1.9979 0.5005 1.7944 2.2246
# policonsumo2.lag 0.9908 1.0093 0.9388 1.0457
# edad_al_ing_1 1.1007 0.9085 1.0904 1.1111
# ano_nac_corr 1.1065 0.9037 1.0963 1.1169
# susinidum_coc_rec2 1.1763 0.8501 0.9849 1.4049
# susinidum_oh 1.0512 0.9513 0.9397 1.1760
# susinidum_pbc 1.0701 0.9345 0.9360 1.2234
# susinidum_mar 1.0842 0.9223 0.9667 1.2160
# psycom_dum_with_rec2 0.9255 1.0805 0.8938 0.9583
# psycom_dum_study 0.8423 1.1873 0.7934 0.8941
# freq_cons_dum_5day 0.9871 1.0131 0.9471 1.0287
# cond_oc_dum_2inact 0.9946 1.0055 0.9400 1.0522
# cond_oc_dum_3unemp 1.0491 0.9532 1.0039 1.0963
# susprindum_oh 1.1321 0.8834 0.9715 1.3191
# susprindum_coc 1.1931 0.8381 1.0205 1.3950
# susprindum_pbc 1.2434 0.8042 1.0657 1.4508
# susprindum_mar 1.2018 0.8321 1.0206 1.4153
#
# Concordance= 0.706 (se = 0.002 )
# Likelihood ratio test= 5371 on 20 df, p=<2e-16
# Wald test = 5053 on 20 df, p=<2e-16
# Score (logrank) test = 5632 on 20 df, p=<2e-16, Robust = 3864 p=<2e-16
#
# (Note: the likelihood ratio and score tests assume independence of
# observations within a cluster, the Wald and robust score tests do not).
summary(iiw_model_after_ph_ba_alt$iiw) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.0643 0.1402 0.5821 1.1179 1.4574 1294.5167 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.0643 0.1402 0.5821 1.1179 1.4574 1294.5167
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("GP intensive ambulatory")
iiw_model_after_ph_gp_ia<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1_rec.lag+
comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod== levels_tipo_de_plan_2_mod[[2]]) %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= subset(maxfu_restr_df, subset= maxfu.id %in% subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod==levels_tipo_de_plan_2_mod[[2]])$id) %>% as.data.frame(),
# formulanull = Surv(time.lag,time,event)~ edad_al_ing_1 +
# ano_nac_corr,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc_rec2", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with_rec2", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome_rec", "comp_bpsc_y3_severe_rec", "less_90d_tr1_rec", "policonsumo2"),
lagfirst= c(2.95082,4.499811/2,0,0,0,0), #90/30.5 4.499811 es 90 días
first= T
)
summary(iiw_model_after_ph_gp_ia$m)Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 15256, number of events= 10260
(1237 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome_rec.lag 0.385527 1.470389 0.059938 0.052500 7.343
log_dias_treat_imp_sin_na.lag -0.159892 0.852236 0.005840 0.012235 -13.068
less_90d_tr1_rec.lag -0.129562 0.878480 0.065487 0.063315 -2.046
comp_bpsc_y3_severe_rec.lag -0.064406 0.937624 0.057456 0.050548 -1.274
policonsumo2.lag -0.751697 0.471566 0.023787 0.023197 -32.405
edad_al_ing_1 0.047348 1.048487 0.004745 0.002975 15.915
ano_nac_corr 0.050791 1.052102 0.004669 0.002920 17.392
susinidum_coc_rec2 0.141599 1.152114 0.073895 0.055690 2.543
susinidum_oh 0.140127 1.150420 0.058018 0.039225 3.572
susinidum_pbc 0.002339 1.002341 0.068236 0.046669 0.050
susinidum_mar 0.169380 1.184570 0.059120 0.040104 4.224
psycom_dum_with_rec2 -0.002221 0.997781 0.019789 0.013096 -0.170
psycom_dum_study -0.285588 0.751572 0.032376 0.027958 -10.215
freq_cons_dum_5day 0.009801 1.009849 0.020311 0.013832 0.709
cond_oc_dum_2inact -0.061440 0.940410 0.030576 0.019679 -3.122
cond_oc_dum_3unemp -0.018074 0.982089 0.021957 0.014496 -1.247
susprindum_oh -0.106767 0.898735 0.087977 0.052865 -2.020
susprindum_coc 0.029272 1.029704 0.088615 0.053023 0.552
susprindum_pbc 0.039693 1.040491 0.086621 0.051892 0.765
susprindum_mar 0.006736 1.006759 0.095242 0.059612 0.113
Pr(>|z|)
tr_outcome_rec.lag 2.08e-13 ***
log_dias_treat_imp_sin_na.lag < 2e-16 ***
less_90d_tr1_rec.lag 0.040727 *
comp_bpsc_y3_severe_rec.lag 0.202613
policonsumo2.lag < 2e-16 ***
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_coc_rec2 0.011003 *
susinidum_oh 0.000354 ***
susinidum_pbc 0.960035
susinidum_mar 2.41e-05 ***
psycom_dum_with_rec2 0.865316
psycom_dum_study < 2e-16 ***
freq_cons_dum_5day 0.478583
cond_oc_dum_2inact 0.001796 **
cond_oc_dum_3unemp 0.212480
susprindum_oh 0.043422 *
susprindum_coc 0.580909
susprindum_pbc 0.444322
susprindum_mar 0.910028
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome_rec.lag 1.4704 0.6801 1.3266 1.6297
log_dias_treat_imp_sin_na.lag 0.8522 1.1734 0.8320 0.8729
less_90d_tr1_rec.lag 0.8785 1.1383 0.7760 0.9945
comp_bpsc_y3_severe_rec.lag 0.9376 1.0665 0.8492 1.0353
policonsumo2.lag 0.4716 2.1206 0.4506 0.4935
edad_al_ing_1 1.0485 0.9538 1.0424 1.0546
ano_nac_corr 1.0521 0.9505 1.0461 1.0581
susinidum_coc_rec2 1.1521 0.8680 1.0330 1.2850
susinidum_oh 1.1504 0.8692 1.0653 1.2424
susinidum_pbc 1.0023 0.9977 0.9147 1.0983
susinidum_mar 1.1846 0.8442 1.0950 1.2814
psycom_dum_with_rec2 0.9978 1.0022 0.9725 1.0237
psycom_dum_study 0.7516 1.3305 0.7115 0.7939
freq_cons_dum_5day 1.0098 0.9902 0.9828 1.0376
cond_oc_dum_2inact 0.9404 1.0634 0.9048 0.9774
cond_oc_dum_3unemp 0.9821 1.0182 0.9546 1.0104
susprindum_oh 0.8987 1.1127 0.8103 0.9969
susprindum_coc 1.0297 0.9712 0.9281 1.1425
susprindum_pbc 1.0405 0.9611 0.9399 1.1519
susprindum_mar 1.0068 0.9933 0.8957 1.1315
Concordance= 0.633 (se = 0.002 )
Likelihood ratio test= 2140 on 20 df, p=<2e-16
Wald test = 2876 on 20 df, p=<2e-16
Score (logrank) test = 2305 on 20 df, p=<2e-16, Robust = 2637 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).Code
# Call:
# coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
# log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
# policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
# susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar, data = datacox, cluster = id)
#
# n= 15256, number of events= 10260
# (1237 observations deleted due to missingness)
#
# coef exp(coef) se(coef) robust se z Pr(>|z|)
# tr_outcome_rec.lag 0.194561 1.214778 0.060033 0.041647 4.672 2.99e-06 ***
# log_dias_treat_imp_sin_na.lag 0.037861 1.038587 0.012193 0.011379 3.327 0.000877 ***
# less_90d_tr1_rec.lag 0.109752 1.116001 0.068389 0.049698 2.208 0.027217 *
# comp_bpsc_y3_severe_rec.lag 0.023291 1.023564 0.058271 0.040852 0.570 0.568590
# policonsumo2.lag -0.865899 0.420673 0.023128 0.019879 -43.559 < 2e-16 ***
# edad_al_ing_1 0.037553 1.038267 0.004720 0.002930 12.819 < 2e-16 ***
# ano_nac_corr 0.042753 1.043681 0.004647 0.002895 14.771 < 2e-16 ***
# susinidum_coc_rec2 0.131486 1.140522 0.074109 0.050774 2.590 0.009607 **
# susinidum_oh 0.145019 1.156062 0.058176 0.037817 3.835 0.000126 ***
# susinidum_pbc -0.007182 0.992844 0.068385 0.045736 -0.157 0.875226
# susinidum_mar 0.175201 1.191485 0.059249 0.038641 4.534 5.79e-06 ***
# psycom_dum_with_rec2 0.019278 1.019465 0.019842 0.011894 1.621 0.105056
# psycom_dum_study -0.242411 0.784734 0.032301 0.026523 -9.140 < 2e-16 ***
# freq_cons_dum_5day 0.018721 1.018897 0.020325 0.013005 1.439 0.150020
# cond_oc_dum_2inact -0.059131 0.942583 0.030582 0.019607 -3.016 0.002562 **
# cond_oc_dum_3unemp 0.008273 1.008307 0.021893 0.013962 0.593 0.553497
# susprindum_oh -0.104867 0.900444 0.088141 0.051923 -2.020 0.043417 *
# susprindum_coc 0.047539 1.048688 0.088742 0.052255 0.910 0.362951
# susprindum_pbc 0.056574 1.058205 0.086757 0.051073 1.108 0.267988
# susprindum_mar 0.058619 1.060371 0.095320 0.056059 1.046 0.295721
# ---
# Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#
# exp(coef) exp(-coef) lower .95 upper .95
# tr_outcome_rec.lag 1.2148 0.8232 1.1196 1.3181
# log_dias_treat_imp_sin_na.lag 1.0386 0.9628 1.0157 1.0620
# less_90d_tr1_rec.lag 1.1160 0.8961 1.0124 1.2302
# comp_bpsc_y3_severe_rec.lag 1.0236 0.9770 0.9448 1.1089
# policonsumo2.lag 0.4207 2.3771 0.4046 0.4374
# edad_al_ing_1 1.0383 0.9631 1.0323 1.0442
# ano_nac_corr 1.0437 0.9581 1.0378 1.0496
# susinidum_coc_rec2 1.1405 0.8768 1.0325 1.2599
# susinidum_oh 1.1561 0.8650 1.0735 1.2450
# susinidum_pbc 0.9928 1.0072 0.9077 1.0860
# susinidum_mar 1.1915 0.8393 1.1046 1.2852
# psycom_dum_with_rec2 1.0195 0.9809 0.9960 1.0435
# psycom_dum_study 0.7847 1.2743 0.7450 0.8266
# freq_cons_dum_5day 1.0189 0.9815 0.9933 1.0452
# cond_oc_dum_2inact 0.9426 1.0609 0.9070 0.9795
# cond_oc_dum_3unemp 1.0083 0.9918 0.9811 1.0363
# susprindum_oh 0.9004 1.1106 0.8133 0.9969
# susprindum_coc 1.0487 0.9536 0.9466 1.1618
# susprindum_pbc 1.0582 0.9450 0.9574 1.1696
# susprindum_mar 1.0604 0.9431 0.9500 1.1835
#
# Concordance= 0.625 (se = 0.002 )
# Likelihood ratio test= 1595 on 20 df, p=<2e-16
# Wald test = 2810 on 20 df, p=<2e-16
# Score (logrank) test = 1706 on 20 df, p=<2e-16, Robust = 2370 p=<2e-16
#
# (Note: the likelihood ratio and score tests assume independence of
# observations within a cluster, the Wald and robust score tests do not).
summary(iiw_model_after_ph_gp_ia$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.2832 0.7168 1.1537 1.4504 2.0894 5.3832 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.5251 0.9191 1.2296 1.4771 1.9988 3.7976
iiw_model_after_ph_gp_ia_alt<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1_rec.lag+
comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod== levels_tipo_de_plan_2_mod[[2]]) %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= subset(maxfu_restr_df, subset= maxfu.id %in% subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod==levels_tipo_de_plan_2_mod[[2]])$id) %>% as.data.frame(),
# formulanull = Surv(time.lag,time,event)~ edad_al_ing_1 +
# ano_nac_corr,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc_rec2", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with_rec2", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome_rec", "comp_bpsc_y3_severe_rec", "less_90d_tr1_rec", "policonsumo2"),
lagfirst= c(2.95082,4.499811,1,1,1,1), #90/30.5 4.499811 es 90 días
first= T
)
summary(iiw_model_after_ph_gp_ia_alt$m)Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 15256, number of events= 10260
(1237 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome_rec.lag 2.014920 7.500129 0.071852 0.068771 29.299
log_dias_treat_imp_sin_na.lag 0.500735 1.649933 0.019681 0.026847 18.652
less_90d_tr1_rec.lag 0.549827 1.732953 0.059743 0.050431 10.903
comp_bpsc_y3_severe_rec.lag 0.433513 1.542667 0.048326 0.045058 9.621
policonsumo2.lag 0.079196 1.082417 0.034122 0.029496 2.685
edad_al_ing_1 0.107546 1.113542 0.004831 0.004325 24.868
ano_nac_corr 0.111431 1.117877 0.004759 0.004253 26.199
susinidum_coc_rec2 0.103013 1.108506 0.072414 0.075162 1.371
susinidum_oh 0.102275 1.107689 0.057331 0.051648 1.980
susinidum_pbc 0.112886 1.119504 0.067849 0.060463 1.867
susinidum_mar 0.157963 1.171122 0.058585 0.052701 2.997
psycom_dum_with_rec2 -0.047860 0.953267 0.019768 0.015955 -3.000
psycom_dum_study -0.281336 0.754775 0.031771 0.031912 -8.816
freq_cons_dum_5day 0.030022 1.030478 0.020274 0.017996 1.668
cond_oc_dum_2inact -0.037308 0.963380 0.030618 0.026755 -1.394
cond_oc_dum_3unemp 0.031426 1.031925 0.021888 0.019230 1.634
susprindum_oh -0.085834 0.917747 0.087841 0.069617 -1.233
susprindum_coc -0.021955 0.978285 0.088512 0.070315 -0.312
susprindum_pbc 0.017908 1.018069 0.086516 0.068698 0.261
susprindum_mar -0.053078 0.948306 0.095181 0.077342 -0.686
Pr(>|z|)
tr_outcome_rec.lag < 2e-16 ***
log_dias_treat_imp_sin_na.lag < 2e-16 ***
less_90d_tr1_rec.lag < 2e-16 ***
comp_bpsc_y3_severe_rec.lag < 2e-16 ***
policonsumo2.lag 0.00725 **
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_coc_rec2 0.17051
susinidum_oh 0.04768 *
susinidum_pbc 0.06190 .
susinidum_mar 0.00272 **
psycom_dum_with_rec2 0.00270 **
psycom_dum_study < 2e-16 ***
freq_cons_dum_5day 0.09525 .
cond_oc_dum_2inact 0.16320
cond_oc_dum_3unemp 0.10222
susprindum_oh 0.21760
susprindum_coc 0.75487
susprindum_pbc 0.79435
susprindum_mar 0.49254
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome_rec.lag 7.5001 0.1333 6.5544 8.5824
log_dias_treat_imp_sin_na.lag 1.6499 0.6061 1.5654 1.7391
less_90d_tr1_rec.lag 1.7330 0.5770 1.5699 1.9130
comp_bpsc_y3_severe_rec.lag 1.5427 0.6482 1.4123 1.6851
policonsumo2.lag 1.0824 0.9239 1.0216 1.1468
edad_al_ing_1 1.1135 0.8980 1.1041 1.1230
ano_nac_corr 1.1179 0.8946 1.1086 1.1272
susinidum_coc_rec2 1.1085 0.9021 0.9567 1.2844
susinidum_oh 1.1077 0.9028 1.0010 1.2257
susinidum_pbc 1.1195 0.8933 0.9944 1.2604
susinidum_mar 1.1711 0.8539 1.0562 1.2986
psycom_dum_with_rec2 0.9533 1.0490 0.9239 0.9835
psycom_dum_study 0.7548 1.3249 0.7090 0.8035
freq_cons_dum_5day 1.0305 0.9704 0.9948 1.0675
cond_oc_dum_2inact 0.9634 1.0380 0.9142 1.0152
cond_oc_dum_3unemp 1.0319 0.9691 0.9938 1.0716
susprindum_oh 0.9177 1.0896 0.8007 1.0519
susprindum_coc 0.9783 1.0222 0.8523 1.1228
susprindum_pbc 1.0181 0.9823 0.8898 1.1648
susprindum_mar 0.9483 1.0545 0.8149 1.1035
Concordance= 0.707 (se = 0.002 )
Likelihood ratio test= 5753 on 20 df, p=<2e-16
Wald test = 5237 on 20 df, p=<2e-16
Score (logrank) test = 5780 on 20 df, p=<2e-16, Robust = 4219 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).Code
# Call:
# coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
# log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
# policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
# susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar, data = datacox, cluster = id)
#
# n= 15256, number of events= 10260
# (1237 observations deleted due to missingness)
#
# coef exp(coef) se(coef) robust se z Pr(>|z|)
# tr_outcome_rec.lag 2.014920 7.500129 0.071852 0.068771 29.299 < 2e-16 ***
# log_dias_treat_imp_sin_na.lag 0.500735 1.649933 0.019681 0.026847 18.652 < 2e-16 ***
# less_90d_tr1_rec.lag 0.549827 1.732953 0.059743 0.050431 10.903 < 2e-16 ***
# comp_bpsc_y3_severe_rec.lag 0.433513 1.542667 0.048326 0.045058 9.621 < 2e-16 ***
# policonsumo2.lag 0.079196 1.082417 0.034122 0.029496 2.685 0.00725 **
# edad_al_ing_1 0.107546 1.113542 0.004831 0.004325 24.868 < 2e-16 ***
# ano_nac_corr 0.111431 1.117877 0.004759 0.004253 26.199 < 2e-16 ***
# susinidum_coc_rec2 0.103013 1.108506 0.072414 0.075162 1.371 0.17051
# susinidum_oh 0.102275 1.107689 0.057331 0.051648 1.980 0.04768 *
# susinidum_pbc 0.112886 1.119504 0.067849 0.060463 1.867 0.06190 .
# susinidum_mar 0.157963 1.171122 0.058585 0.052701 2.997 0.00272 **
# psycom_dum_with_rec2 -0.047860 0.953267 0.019768 0.015955 -3.000 0.00270 **
# psycom_dum_study -0.281336 0.754775 0.031771 0.031912 -8.816 < 2e-16 ***
# freq_cons_dum_5day 0.030022 1.030478 0.020274 0.017996 1.668 0.09525 .
# cond_oc_dum_2inact -0.037308 0.963380 0.030618 0.026755 -1.394 0.16320
# cond_oc_dum_3unemp 0.031426 1.031925 0.021888 0.019230 1.634 0.10222
# susprindum_oh -0.085834 0.917747 0.087841 0.069617 -1.233 0.21760
# susprindum_coc -0.021955 0.978285 0.088512 0.070315 -0.312 0.75487
# susprindum_pbc 0.017908 1.018069 0.086516 0.068698 0.261 0.79435
# susprindum_mar -0.053078 0.948306 0.095181 0.077342 -0.686 0.49254
# ---
# Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#
# exp(coef) exp(-coef) lower .95 upper .95
# tr_outcome_rec.lag 7.5001 0.1333 6.5544 8.5824
# log_dias_treat_imp_sin_na.lag 1.6499 0.6061 1.5654 1.7391
# less_90d_tr1_rec.lag 1.7330 0.5770 1.5699 1.9130
# comp_bpsc_y3_severe_rec.lag 1.5427 0.6482 1.4123 1.6851
# policonsumo2.lag 1.0824 0.9239 1.0216 1.1468
# edad_al_ing_1 1.1135 0.8980 1.1041 1.1230
# ano_nac_corr 1.1179 0.8946 1.1086 1.1272
# susinidum_coc_rec2 1.1085 0.9021 0.9567 1.2844
# susinidum_oh 1.1077 0.9028 1.0010 1.2257
# susinidum_pbc 1.1195 0.8933 0.9944 1.2604
# susinidum_mar 1.1711 0.8539 1.0562 1.2986
# psycom_dum_with_rec2 0.9533 1.0490 0.9239 0.9835
# psycom_dum_study 0.7548 1.3249 0.7090 0.8035
# freq_cons_dum_5day 1.0305 0.9704 0.9948 1.0675
# cond_oc_dum_2inact 0.9634 1.0380 0.9142 1.0152
# cond_oc_dum_3unemp 1.0319 0.9691 0.9938 1.0716
# susprindum_oh 0.9177 1.0896 0.8007 1.0519
# susprindum_coc 0.9783 1.0222 0.8523 1.1228
# susprindum_pbc 1.0181 0.9823 0.8898 1.1648
# susprindum_mar 0.9483 1.0545 0.8149 1.1035
#
# Concordance= 0.707 (se = 0.002 )
# Likelihood ratio test= 5753 on 20 df, p=<2e-16
# Wald test = 5237 on 20 df, p=<2e-16
# Score (logrank) test = 5780 on 20 df, p=<2e-16, Robust = 4219 p=<2e-16
#
# (Note: the likelihood ratio and score tests assume independence of
# observations within a cluster, the Wald and robust score tests do not).
summary(iiw_model_after_ph_gp_ia_alt$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.0678 0.1663 0.5251 1.4332 1.3754 1183.8688 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.0678 0.1663 0.5251 1.4332 1.3754 1183.8688
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("GP residential")
iiw_model_after_ph_gp_res<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1_rec.lag+
comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod== levels_tipo_de_plan_2_mod[[3]]) %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= subset(maxfu_restr_df, subset= maxfu.id %in% subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod==levels_tipo_de_plan_2_mod[[3]])$id) %>% as.data.frame(),
# formulanull = Surv(time.lag,time,event)~ edad_al_ing_1 +
# ano_nac_corr,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc_rec2", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with_rec2", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome_rec", "comp_bpsc_y3_severe_rec", "less_90d_tr1_rec", "policonsumo2"),
lagfirst= c(2.95082,4.499811/2,0,0,0,0), #90/30.5 4.499811 es 90 días
first= T
)
summary(iiw_model_after_ph_gp_res$m)Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 6745, number of events= 4567
(637 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome_rec.lag 0.128188 1.136767 0.101990 0.086925 1.475
log_dias_treat_imp_sin_na.lag -0.085496 0.918057 0.008798 0.011288 -7.574
less_90d_tr1_rec.lag -0.088387 0.915407 0.103118 0.084266 -1.049
comp_bpsc_y3_severe_rec.lag 0.175771 1.192165 0.082749 0.081533 2.156
policonsumo2.lag -0.750680 0.472045 0.036058 0.034731 -21.614
edad_al_ing_1 0.028939 1.029361 0.007348 0.004632 6.248
ano_nac_corr 0.032926 1.033474 0.007304 0.004592 7.170
susinidum_coc_rec2 0.183915 1.201913 0.125673 0.096940 1.897
susinidum_oh 0.138489 1.148536 0.094577 0.076211 1.817
susinidum_pbc 0.032443 1.032975 0.101841 0.081788 0.397
susinidum_mar 0.118222 1.125494 0.094812 0.076811 1.539
psycom_dum_with_rec2 0.056826 1.058472 0.029195 0.019809 2.869
psycom_dum_study -0.207341 0.812742 0.048764 0.037096 -5.589
freq_cons_dum_5day -0.027696 0.972684 0.034087 0.023537 -1.177
cond_oc_dum_2inact 0.091545 1.095866 0.055552 0.040162 2.279
cond_oc_dum_3unemp 0.107838 1.113867 0.045740 0.031919 3.379
susprindum_oh -0.058676 0.943012 0.158396 0.098428 -0.596
susprindum_coc 0.091597 1.095923 0.159134 0.098627 0.929
susprindum_pbc 0.039675 1.040473 0.154505 0.094983 0.418
susprindum_mar -0.075467 0.927310 0.178402 0.113375 -0.666
Pr(>|z|)
tr_outcome_rec.lag 0.140295
log_dias_treat_imp_sin_na.lag 3.62e-14 ***
less_90d_tr1_rec.lag 0.294224
comp_bpsc_y3_severe_rec.lag 0.031096 *
policonsumo2.lag < 2e-16 ***
edad_al_ing_1 4.17e-10 ***
ano_nac_corr 7.52e-13 ***
susinidum_coc_rec2 0.057802 .
susinidum_oh 0.069192 .
susinidum_pbc 0.691608
susinidum_mar 0.123774
psycom_dum_with_rec2 0.004121 **
psycom_dum_study 2.28e-08 ***
freq_cons_dum_5day 0.239318
cond_oc_dum_2inact 0.022644 *
cond_oc_dum_3unemp 0.000729 ***
susprindum_oh 0.551086
susprindum_coc 0.353037
susprindum_pbc 0.676161
susprindum_mar 0.505641
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome_rec.lag 1.1368 0.8797 0.9587 1.3479
log_dias_treat_imp_sin_na.lag 0.9181 1.0893 0.8980 0.9386
less_90d_tr1_rec.lag 0.9154 1.0924 0.7760 1.0798
comp_bpsc_y3_severe_rec.lag 1.1922 0.8388 1.0161 1.3987
policonsumo2.lag 0.4720 2.1184 0.4410 0.5053
edad_al_ing_1 1.0294 0.9715 1.0201 1.0387
ano_nac_corr 1.0335 0.9676 1.0242 1.0428
susinidum_coc_rec2 1.2019 0.8320 0.9939 1.4534
susinidum_oh 1.1485 0.8707 0.9892 1.3336
susinidum_pbc 1.0330 0.9681 0.8800 1.2126
susinidum_mar 1.1255 0.8885 0.9682 1.3084
psycom_dum_with_rec2 1.0585 0.9448 1.0182 1.1004
psycom_dum_study 0.8127 1.2304 0.7557 0.8740
freq_cons_dum_5day 0.9727 1.0281 0.9288 1.0186
cond_oc_dum_2inact 1.0959 0.9125 1.0129 1.1856
cond_oc_dum_3unemp 1.1139 0.8978 1.0463 1.1858
susprindum_oh 0.9430 1.0604 0.7776 1.1437
susprindum_coc 1.0959 0.9125 0.9033 1.3296
susprindum_pbc 1.0405 0.9611 0.8637 1.2534
susprindum_mar 0.9273 1.0784 0.7425 1.1581
Concordance= 0.605 (se = 0.003 )
Likelihood ratio test= 695.9 on 20 df, p=<2e-16
Wald test = 1085 on 20 df, p=<2e-16
Score (logrank) test = 743.2 on 20 df, p=<2e-16, Robust = 996.3 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).Code
# Call:
# coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
# log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
# policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
# susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar, data = datacox, cluster = id)
#
# n= 6745, number of events= 4567
# (637 observations deleted due to missingness)
#
# coef exp(coef) se(coef) robust se z Pr(>|z|)
# tr_outcome_rec.lag -0.039009 0.961742 0.105701 0.074349 -0.525 0.599808
# log_dias_treat_imp_sin_na.lag 0.078977 1.082180 0.015647 0.013905 5.680 1.35e-08 ***
# less_90d_tr1_rec.lag 0.150856 1.162829 0.107295 0.080559 1.873 0.061122 .
# comp_bpsc_y3_severe_rec.lag 0.353959 1.424696 0.090172 0.070675 5.008 5.49e-07 ***
# policonsumo2.lag -0.837956 0.432594 0.035119 0.032878 -25.487 < 2e-16 ***
# edad_al_ing_1 0.025965 1.026305 0.007319 0.004690 5.536 3.09e-08 ***
# ano_nac_corr 0.030994 1.031479 0.007283 0.004684 6.617 3.66e-11 ***
# susinidum_coc_rec2 0.160851 1.174510 0.124649 0.094183 1.708 0.087664 .
# susinidum_oh 0.114160 1.120931 0.094181 0.074294 1.537 0.124391
# susinidum_pbc 0.009529 1.009575 0.101467 0.080150 0.119 0.905358
# susinidum_mar 0.103063 1.108562 0.094427 0.074787 1.378 0.168173
# psycom_dum_with_rec2 0.071582 1.074206 0.029128 0.019756 3.623 0.000291 ***
# psycom_dum_study -0.187465 0.829058 0.048601 0.037539 -4.994 5.92e-07 ***
# freq_cons_dum_5day -0.017574 0.982579 0.034084 0.023154 -0.759 0.447831
# cond_oc_dum_2inact 0.102551 1.107994 0.055537 0.038242 2.682 0.007327 **
# cond_oc_dum_3unemp 0.099801 1.104951 0.045743 0.031127 3.206 0.001345 **
# susprindum_oh -0.030306 0.970149 0.158516 0.107214 -0.283 0.777431
# susprindum_coc 0.111599 1.118064 0.159135 0.107984 1.033 0.301381
# susprindum_pbc 0.080561 1.083895 0.154488 0.104564 0.770 0.441033
# susprindum_mar 0.064293 1.066405 0.177941 0.113070 0.569 0.569619
# ---
# Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#
# exp(coef) exp(-coef) lower .95 upper .95
# tr_outcome_rec.lag 0.9617 1.0398 0.8313 1.1126
# log_dias_treat_imp_sin_na.lag 1.0822 0.9241 1.0531 1.1121
# less_90d_tr1_rec.lag 1.1628 0.8600 0.9930 1.3617
# comp_bpsc_y3_severe_rec.lag 1.4247 0.7019 1.2404 1.6364
# policonsumo2.lag 0.4326 2.3116 0.4056 0.4614
# edad_al_ing_1 1.0263 0.9744 1.0169 1.0358
# ano_nac_corr 1.0315 0.9695 1.0221 1.0410
# susinidum_coc_rec2 1.1745 0.8514 0.9765 1.4126
# susinidum_oh 1.1209 0.8921 0.9690 1.2966
# susinidum_pbc 1.0096 0.9905 0.8628 1.1813
# susinidum_mar 1.1086 0.9021 0.9574 1.2836
# psycom_dum_with_rec2 1.0742 0.9309 1.0334 1.1166
# psycom_dum_study 0.8291 1.2062 0.7702 0.8924
# freq_cons_dum_5day 0.9826 1.0177 0.9390 1.0282
# cond_oc_dum_2inact 1.1080 0.9025 1.0280 1.1942
# cond_oc_dum_3unemp 1.1050 0.9050 1.0396 1.1745
# susprindum_oh 0.9701 1.0308 0.7863 1.1970
# susprindum_coc 1.1181 0.8944 0.9048 1.3816
# susprindum_pbc 1.0839 0.9226 0.8830 1.3304
# susprindum_mar 1.0664 0.9377 0.8544 1.3310
#
# Concordance= 0.618 (se = 0.003 )
# Likelihood ratio test= 644.2 on 20 df, p=<2e-16
# Wald test = 1048 on 20 df, p=<2e-16
# Score (logrank) test = 689.3 on 20 df, p=<2e-16, Robust = 950.3 p=<2e-16
#
# (Note: the likelihood ratio and score tests assume independence of
# observations within a cluster, the Wald and robust score tests do not).
summary(iiw_model_after_ph_gp_res$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.3321 0.7511 1.1708 1.3145 1.8355 3.2183 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.5131 0.8729 1.2617 1.3565 1.7689 6.7300
iiw_model_after_ph_gp_res_alt<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1_rec.lag+
comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod== levels_tipo_de_plan_2_mod[[3]]) %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= subset(maxfu_restr_df, subset= maxfu.id %in% subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod==levels_tipo_de_plan_2_mod[[3]])$id) %>% as.data.frame(),
# formulanull = Surv(time.lag,time,event)~ edad_al_ing_1 +
# ano_nac_corr,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc_rec2", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with_rec2", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome_rec", "comp_bpsc_y3_severe_rec", "less_90d_tr1_rec", "policonsumo2"),
lagfirst= c(2.95082,4.499811,1,1,1,1), #90/30.5 4.499811 es 90 días
first= T
)
summary(iiw_model_after_ph_gp_res_alt$m)Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 6745, number of events= 4567
(637 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome_rec.lag 0.981483 2.668410 0.120321 0.103018 9.527
log_dias_treat_imp_sin_na.lag 0.328445 1.388807 0.021738 0.031862 10.308
less_90d_tr1_rec.lag 0.482546 1.620195 0.097638 0.084879 5.685
comp_bpsc_y3_severe_rec.lag 0.940955 2.562426 0.100147 0.088610 10.619
policonsumo2.lag 0.161177 1.174893 0.054428 0.050466 3.194
edad_al_ing_1 0.060548 1.062418 0.007416 0.006446 9.393
ano_nac_corr 0.063469 1.065526 0.007369 0.006388 9.936
susinidum_coc_rec2 0.226938 1.254752 0.122208 0.130113 1.744
susinidum_oh 0.083881 1.087500 0.093000 0.086277 0.972
susinidum_pbc 0.163264 1.177348 0.100532 0.094265 1.732
susinidum_mar 0.107515 1.113507 0.093363 0.086830 1.238
psycom_dum_with_rec2 -0.050914 0.950360 0.029484 0.024602 -2.069
psycom_dum_study -0.357490 0.699430 0.047553 0.046607 -7.670
freq_cons_dum_5day -0.012423 0.987654 0.033856 0.029387 -0.423
cond_oc_dum_2inact 0.106389 1.112255 0.055170 0.048125 2.211
cond_oc_dum_3unemp 0.065295 1.067474 0.045224 0.039172 1.667
susprindum_oh 0.101693 1.107044 0.158048 0.131136 0.775
susprindum_coc 0.046216 1.047300 0.159165 0.131835 0.351
susprindum_pbc 0.015162 1.015277 0.154594 0.127081 0.119
susprindum_mar 0.042951 1.043887 0.177913 0.138399 0.310
Pr(>|z|)
tr_outcome_rec.lag < 2e-16 ***
log_dias_treat_imp_sin_na.lag < 2e-16 ***
less_90d_tr1_rec.lag 1.31e-08 ***
comp_bpsc_y3_severe_rec.lag < 2e-16 ***
policonsumo2.lag 0.0014 **
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_coc_rec2 0.0811 .
susinidum_oh 0.3309
susinidum_pbc 0.0833 .
susinidum_mar 0.2156
psycom_dum_with_rec2 0.0385 *
psycom_dum_study 1.72e-14 ***
freq_cons_dum_5day 0.6725
cond_oc_dum_2inact 0.0271 *
cond_oc_dum_3unemp 0.0955 .
susprindum_oh 0.4381
susprindum_coc 0.7259
susprindum_pbc 0.9050
susprindum_mar 0.7563
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome_rec.lag 2.6684 0.3748 2.1805 3.2654
log_dias_treat_imp_sin_na.lag 1.3888 0.7200 1.3047 1.4783
less_90d_tr1_rec.lag 1.6202 0.6172 1.3719 1.9134
comp_bpsc_y3_severe_rec.lag 2.5624 0.3903 2.1539 3.0484
policonsumo2.lag 1.1749 0.8511 1.0642 1.2970
edad_al_ing_1 1.0624 0.9412 1.0491 1.0759
ano_nac_corr 1.0655 0.9385 1.0523 1.0790
susinidum_coc_rec2 1.2548 0.7970 0.9723 1.6192
susinidum_oh 1.0875 0.9195 0.9183 1.2879
susinidum_pbc 1.1773 0.8494 0.9787 1.4163
susinidum_mar 1.1135 0.8981 0.9393 1.3201
psycom_dum_with_rec2 0.9504 1.0522 0.9056 0.9973
psycom_dum_study 0.6994 1.4297 0.6384 0.7663
freq_cons_dum_5day 0.9877 1.0125 0.9324 1.0462
cond_oc_dum_2inact 1.1123 0.8991 1.0121 1.2223
cond_oc_dum_3unemp 1.0675 0.9368 0.9886 1.1527
susprindum_oh 1.1070 0.9033 0.8561 1.4315
susprindum_coc 1.0473 0.9548 0.8088 1.3561
susprindum_pbc 1.0153 0.9850 0.7914 1.3024
susprindum_mar 1.0439 0.9580 0.7959 1.3692
Concordance= 0.682 (se = 0.003 )
Likelihood ratio test= 1994 on 20 df, p=<2e-16
Wald test = 2159 on 20 df, p=<2e-16
Score (logrank) test = 2003 on 20 df, p=<2e-16, Robust = 1726 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).Code
# Call:
# coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
# log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
# policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
# susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar, data = datacox, cluster = id)
#
# n= 6745, number of events= 4567
# (637 observations deleted due to missingness)
#
# coef exp(coef) se(coef) robust se z Pr(>|z|)
# tr_outcome_rec.lag 0.981483 2.668410 0.120321 0.103018 9.527 < 2e-16 ***
# log_dias_treat_imp_sin_na.lag 0.328445 1.388807 0.021738 0.031862 10.308 < 2e-16 ***
# less_90d_tr1_rec.lag 0.482546 1.620195 0.097638 0.084879 5.685 1.31e-08 ***
# comp_bpsc_y3_severe_rec.lag 0.940955 2.562426 0.100147 0.088610 10.619 < 2e-16 ***
# policonsumo2.lag 0.161177 1.174893 0.054428 0.050466 3.194 0.0014 **
# edad_al_ing_1 0.060548 1.062418 0.007416 0.006446 9.393 < 2e-16 ***
# ano_nac_corr 0.063469 1.065526 0.007369 0.006388 9.936 < 2e-16 ***
# susinidum_coc_rec2 0.226938 1.254752 0.122208 0.130113 1.744 0.0811 .
# susinidum_oh 0.083881 1.087500 0.093000 0.086277 0.972 0.3309
# susinidum_pbc 0.163264 1.177348 0.100532 0.094265 1.732 0.0833 .
# susinidum_mar 0.107515 1.113507 0.093363 0.086830 1.238 0.2156
# psycom_dum_with_rec2 -0.050914 0.950360 0.029484 0.024602 -2.069 0.0385 *
# psycom_dum_study -0.357490 0.699430 0.047553 0.046607 -7.670 1.72e-14 ***
# freq_cons_dum_5day -0.012423 0.987654 0.033856 0.029387 -0.423 0.6725
# cond_oc_dum_2inact 0.106389 1.112255 0.055170 0.048125 2.211 0.0271 *
# cond_oc_dum_3unemp 0.065295 1.067474 0.045224 0.039172 1.667 0.0955 .
# susprindum_oh 0.101693 1.107044 0.158048 0.131136 0.775 0.4381
# susprindum_coc 0.046216 1.047300 0.159165 0.131835 0.351 0.7259
# susprindum_pbc 0.015162 1.015277 0.154594 0.127081 0.119 0.9050
# susprindum_mar 0.042951 1.043887 0.177913 0.138399 0.310 0.7563
# ---
# Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#
# exp(coef) exp(-coef) lower .95 upper .95
# tr_outcome_rec.lag 2.6684 0.3748 2.1805 3.2654
# log_dias_treat_imp_sin_na.lag 1.3888 0.7200 1.3047 1.4783
# less_90d_tr1_rec.lag 1.6202 0.6172 1.3719 1.9134
# comp_bpsc_y3_severe_rec.lag 2.5624 0.3903 2.1539 3.0484
# policonsumo2.lag 1.1749 0.8511 1.0642 1.2970
# edad_al_ing_1 1.0624 0.9412 1.0491 1.0759
# ano_nac_corr 1.0655 0.9385 1.0523 1.0790
# susinidum_coc_rec2 1.2548 0.7970 0.9723 1.6192
# susinidum_oh 1.0875 0.9195 0.9183 1.2879
# susinidum_pbc 1.1773 0.8494 0.9787 1.4163
# susinidum_mar 1.1135 0.8981 0.9393 1.3201
# psycom_dum_with_rec2 0.9504 1.0522 0.9056 0.9973
# psycom_dum_study 0.6994 1.4297 0.6384 0.7663
# freq_cons_dum_5day 0.9877 1.0125 0.9324 1.0462
# cond_oc_dum_2inact 1.1123 0.8991 1.0121 1.2223
# cond_oc_dum_3unemp 1.0675 0.9368 0.9886 1.1527
# susprindum_oh 1.1070 0.9033 0.8561 1.4315
# susprindum_coc 1.0473 0.9548 0.8088 1.3561
# susprindum_pbc 1.0153 0.9850 0.7914 1.3024
# susprindum_mar 1.0439 0.9580 0.7959 1.3692
#
# Concordance= 0.682 (se = 0.003 )
# Likelihood ratio test= 1994 on 20 df, p=<2e-16
# Wald test = 2159 on 20 df, p=<2e-16
# Score (logrank) test = 2003 on 20 df, p=<2e-16, Robust = 1726 p=<2e-16
#
# (Note: the likelihood ratio and score tests assume independence of
# observations within a cluster, the Wald and robust score tests do not).
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("WO intensive ambulatory")
iiw_model_after_ph_wo_ia<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1_rec.lag+
comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod== levels_tipo_de_plan_2_mod[[4]]) %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= subset(maxfu_restr_df, subset= maxfu.id %in% subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod==levels_tipo_de_plan_2_mod[[4]])$id) %>% as.data.frame(),
# formulanull = Surv(time.lag,time,event)~ edad_al_ing_1 +
# ano_nac_corr,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc_rec2", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with_rec2", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome_rec", "comp_bpsc_y3_severe_rec", "less_90d_tr1_rec", "policonsumo2"),
lagfirst= c(2.95082,4.499811/2,0,0,0,0), #90/30.5 4.499811 es 90 días
first= T
)
summary(iiw_model_after_ph_wo_ia$m)Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 2342, number of events= 1597
(163 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome_rec.lag 0.590396 1.804703 0.165929 0.118174 4.996
log_dias_treat_imp_sin_na.lag -0.243294 0.784041 0.022535 0.025535 -9.528
less_90d_tr1_rec.lag -0.206354 0.813545 0.194073 0.151714 -1.360
comp_bpsc_y3_severe_rec.lag -0.132974 0.875488 0.164473 0.115583 -1.150
policonsumo2.lag -0.615544 0.540347 0.061535 0.054363 -11.323
edad_al_ing_1 0.037924 1.038652 0.012119 0.007974 4.756
ano_nac_corr 0.042599 1.043519 0.011695 0.007815 5.451
susinidum_coc_rec2 -0.042127 0.958748 0.170050 0.135358 -0.311
susinidum_oh -0.025163 0.975151 0.143634 0.096426 -0.261
susinidum_pbc -0.020822 0.979393 0.156236 0.108823 -0.191
susinidum_mar 0.012483 1.012562 0.144243 0.094129 0.133
psycom_dum_with_rec2 0.016660 1.016799 0.049880 0.031631 0.527
psycom_dum_study -0.275828 0.758943 0.085551 0.072534 -3.803
freq_cons_dum_5day 0.031700 1.032208 0.051667 0.032640 0.971
cond_oc_dum_2inact -0.018931 0.981247 0.065151 0.042881 -0.441
cond_oc_dum_3unemp -0.005474 0.994541 0.067826 0.042655 -0.128
susprindum_oh 0.088216 1.092224 0.183762 0.117491 0.751
susprindum_coc 0.128627 1.137265 0.186221 0.126652 1.016
susprindum_pbc 0.100929 1.106199 0.180469 0.123526 0.817
susprindum_mar 0.086862 1.090746 0.199405 0.140061 0.620
Pr(>|z|)
tr_outcome_rec.lag 5.85e-07 ***
log_dias_treat_imp_sin_na.lag < 2e-16 ***
less_90d_tr1_rec.lag 0.173783
comp_bpsc_y3_severe_rec.lag 0.249954
policonsumo2.lag < 2e-16 ***
edad_al_ing_1 1.97e-06 ***
ano_nac_corr 5.00e-08 ***
susinidum_coc_rec2 0.755627
susinidum_oh 0.794129
susinidum_pbc 0.848260
susinidum_mar 0.894494
psycom_dum_with_rec2 0.598406
psycom_dum_study 0.000143 ***
freq_cons_dum_5day 0.331444
cond_oc_dum_2inact 0.658865
cond_oc_dum_3unemp 0.897892
susprindum_oh 0.452755
susprindum_coc 0.309823
susprindum_pbc 0.413890
susprindum_mar 0.535147
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome_rec.lag 1.8047 0.5541 1.4316 2.2751
log_dias_treat_imp_sin_na.lag 0.7840 1.2754 0.7458 0.8243
less_90d_tr1_rec.lag 0.8135 1.2292 0.6043 1.0953
comp_bpsc_y3_severe_rec.lag 0.8755 1.1422 0.6980 1.0981
policonsumo2.lag 0.5403 1.8507 0.4857 0.6011
edad_al_ing_1 1.0387 0.9628 1.0225 1.0550
ano_nac_corr 1.0435 0.9583 1.0277 1.0596
susinidum_coc_rec2 0.9587 1.0430 0.7353 1.2500
susinidum_oh 0.9752 1.0255 0.8072 1.1780
susinidum_pbc 0.9794 1.0210 0.7913 1.2122
susinidum_mar 1.0126 0.9876 0.8420 1.2177
psycom_dum_with_rec2 1.0168 0.9835 0.9557 1.0818
psycom_dum_study 0.7589 1.3176 0.6584 0.8749
freq_cons_dum_5day 1.0322 0.9688 0.9682 1.1004
cond_oc_dum_2inact 0.9812 1.0191 0.9021 1.0673
cond_oc_dum_3unemp 0.9945 1.0055 0.9148 1.0813
susprindum_oh 1.0922 0.9156 0.8676 1.3751
susprindum_coc 1.1373 0.8793 0.8873 1.4577
susprindum_pbc 1.1062 0.9040 0.8683 1.4092
susprindum_mar 1.0907 0.9168 0.8289 1.4353
Concordance= 0.634 (se = 0.006 )
Likelihood ratio test= 334.3 on 20 df, p=<2e-16
Wald test = 467.2 on 20 df, p=<2e-16
Score (logrank) test = 346.7 on 20 df, p=<2e-16, Robust = 401.6 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).Code
# Call:
# coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
# log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
# policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
# susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar, data = datacox, cluster = id)
#
# n= 2342, number of events= 1597
# (163 observations deleted due to missingness)
#
# coef exp(coef) se(coef) robust se z Pr(>|z|)
# tr_outcome_rec.lag 0.0960769 1.1008437 0.1608893 0.1073787 0.895 0.370922
# log_dias_treat_imp_sin_na.lag 0.0601846 1.0620325 0.0341951 0.0323800 1.859 0.063070 .
# less_90d_tr1_rec.lag 0.3920119 1.4799554 0.1955210 0.1410453 2.779 0.005447 **
# comp_bpsc_y3_severe_rec.lag 0.0021099 1.0021122 0.1634728 0.1074412 0.020 0.984332
# policonsumo2.lag -0.8194113 0.4406910 0.0577374 0.0471515 -17.378 < 2e-16 ***
# edad_al_ing_1 0.0264262 1.0267785 0.0120769 0.0079482 3.325 0.000885 ***
# ano_nac_corr 0.0326238 1.0331618 0.0116686 0.0077493 4.210 2.55e-05 ***
# susinidum_coc_rec2 0.0065845 1.0066062 0.1708421 0.1250574 0.053 0.958010
# susinidum_oh 0.0354693 1.0361059 0.1442859 0.1006518 0.352 0.724541
# susinidum_pbc 0.0354305 1.0360656 0.1568797 0.1101749 0.322 0.747768
# susinidum_mar 0.0902561 1.0944546 0.1448738 0.0980071 0.921 0.357095
# psycom_dum_with_rec2 0.0288659 1.0292865 0.0500612 0.0300631 0.960 0.336967
# psycom_dum_study -0.2367838 0.7891619 0.0863122 0.0688161 -3.441 0.000580 ***
# freq_cons_dum_5day 0.0330127 1.0335637 0.0517079 0.0316338 1.044 0.296676
# cond_oc_dum_2inact -0.0004428 0.9995573 0.0653684 0.0422072 -0.010 0.991630
# cond_oc_dum_3unemp -0.0078465 0.9921842 0.0678832 0.0425011 -0.185 0.853528
# susprindum_oh 0.0346128 1.0352188 0.1838352 0.1319090 0.262 0.793014
# susprindum_coc 0.1080539 1.1141078 0.1862425 0.1412067 0.765 0.444142
# susprindum_pbc 0.1142682 1.1210528 0.1808628 0.1382042 0.827 0.408346
# susprindum_mar 0.1072029 1.1131601 0.1989858 0.1539935 0.696 0.486334
# ---
# Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#
# exp(coef) exp(-coef) lower .95 upper .95
# tr_outcome_rec.lag 1.1008 0.9084 0.8919 1.3587
# log_dias_treat_imp_sin_na.lag 1.0620 0.9416 0.9967 1.1316
# less_90d_tr1_rec.lag 1.4800 0.6757 1.1225 1.9512
# comp_bpsc_y3_severe_rec.lag 1.0021 0.9979 0.8118 1.2370
# policonsumo2.lag 0.4407 2.2692 0.4018 0.4834
# edad_al_ing_1 1.0268 0.9739 1.0109 1.0429
# ano_nac_corr 1.0332 0.9679 1.0176 1.0490
# susinidum_coc_rec2 1.0066 0.9934 0.7878 1.2862
# susinidum_oh 1.0361 0.9652 0.8506 1.2621
# susinidum_pbc 1.0361 0.9652 0.8348 1.2858
# susinidum_mar 1.0945 0.9137 0.9032 1.3262
# psycom_dum_with_rec2 1.0293 0.9715 0.9704 1.0918
# psycom_dum_study 0.7892 1.2672 0.6896 0.9031
# freq_cons_dum_5day 1.0336 0.9675 0.9714 1.0997
# cond_oc_dum_2inact 0.9996 1.0004 0.9202 1.0858
# cond_oc_dum_3unemp 0.9922 1.0079 0.9129 1.0784
# susprindum_oh 1.0352 0.9660 0.7994 1.3406
# susprindum_coc 1.1141 0.8976 0.8448 1.4693
# susprindum_pbc 1.1211 0.8920 0.8550 1.4698
# susprindum_mar 1.1132 0.8983 0.8231 1.5053
#
# Concordance= 0.627 (se = 0.005 )
# Likelihood ratio test= 226.7 on 20 df, p=<2e-16
# Wald test = 427.2 on 20 df, p=<2e-16
# Score (logrank) test = 241.2 on 20 df, p=<2e-16, Robust = 343 p=<2e-16
#
# (Note: the likelihood ratio and score tests assume independence of
# observations within a cluster, the Wald and robust score tests do not).
summary(iiw_model_after_ph_wo_ia$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.336 0.616 1.124 1.368 1.966 5.427 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.5302 0.9069 1.1501 1.3870 1.8680 3.0902
iiw_model_after_ph_wo_ia_alt<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1_rec.lag+
comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod== levels_tipo_de_plan_2_mod[[4]]) %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= subset(maxfu_restr_df, subset= maxfu.id %in% subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod==levels_tipo_de_plan_2_mod[[4]])$id) %>% as.data.frame(),
# formulanull = Surv(time.lag,time,event)~ edad_al_ing_1 +
# ano_nac_corr,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc_rec2", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with_rec2", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome_rec", "comp_bpsc_y3_severe_rec", "less_90d_tr1_rec", "policonsumo2"),
lagfirst= c(2.95082,4.499811,1,1,1,1), #90/30.5 4.499811 es 90 días
first= T
)
summary(iiw_model_after_ph_wo_ia_alt$m)Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 2342, number of events= 1597
(163 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome_rec.lag 2.07668 7.97793 0.18951 0.17050 12.180
log_dias_treat_imp_sin_na.lag 0.48119 1.61800 0.04878 0.05862 8.208
less_90d_tr1_rec.lag 0.53253 1.70324 0.15975 0.12591 4.230
comp_bpsc_y3_severe_rec.lag 0.24739 1.28067 0.13099 0.12212 2.026
policonsumo2.lag -0.06660 0.93557 0.07911 0.07156 -0.931
edad_al_ing_1 0.10711 1.11305 0.01271 0.01175 9.116
ano_nac_corr 0.11118 1.11759 0.01230 0.01146 9.705
susinidum_coc_rec2 -0.19912 0.81945 0.16603 0.16462 -1.210
susinidum_oh -0.24978 0.77897 0.14191 0.12065 -2.070
susinidum_pbc -0.07015 0.93225 0.15381 0.13115 -0.535
susinidum_mar -0.14500 0.86502 0.14235 0.11908 -1.218
psycom_dum_with_rec2 -0.03390 0.96667 0.05027 0.04078 -0.831
psycom_dum_study -0.22053 0.80210 0.08298 0.08202 -2.689
freq_cons_dum_5day 0.03082 1.03130 0.05150 0.04617 0.668
cond_oc_dum_2inact 0.02049 1.02070 0.06522 0.06013 0.341
cond_oc_dum_3unemp -0.01038 0.98968 0.06734 0.06267 -0.166
susprindum_oh 0.03448 1.03508 0.18281 0.17658 0.195
susprindum_coc 0.06938 1.07184 0.18629 0.18800 0.369
susprindum_pbc 0.11002 1.11631 0.18066 0.18587 0.592
susprindum_mar 0.05398 1.05546 0.19901 0.20360 0.265
Pr(>|z|)
tr_outcome_rec.lag < 2e-16 ***
log_dias_treat_imp_sin_na.lag 2.24e-16 ***
less_90d_tr1_rec.lag 2.34e-05 ***
comp_bpsc_y3_severe_rec.lag 0.04280 *
policonsumo2.lag 0.35204
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_coc_rec2 0.22642
susinidum_oh 0.03842 *
susinidum_pbc 0.59273
susinidum_mar 0.22334
psycom_dum_with_rec2 0.40585
psycom_dum_study 0.00717 **
freq_cons_dum_5day 0.50442
cond_oc_dum_2inact 0.73326
cond_oc_dum_3unemp 0.86850
susprindum_oh 0.84517
susprindum_coc 0.71210
susprindum_pbc 0.55390
susprindum_mar 0.79092
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome_rec.lag 7.9779 0.1253 5.7117 11.1434
log_dias_treat_imp_sin_na.lag 1.6180 0.6180 1.4424 1.8150
less_90d_tr1_rec.lag 1.7032 0.5871 1.3308 2.1799
comp_bpsc_y3_severe_rec.lag 1.2807 0.7808 1.0081 1.6270
policonsumo2.lag 0.9356 1.0689 0.8131 1.0764
edad_al_ing_1 1.1131 0.8984 1.0877 1.1390
ano_nac_corr 1.1176 0.8948 1.0928 1.1430
susinidum_coc_rec2 0.8194 1.2203 0.5935 1.1315
susinidum_oh 0.7790 1.2837 0.6149 0.9868
susinidum_pbc 0.9323 1.0727 0.7209 1.2055
susinidum_mar 0.8650 1.1560 0.6850 1.0924
psycom_dum_with_rec2 0.9667 1.0345 0.8924 1.0471
psycom_dum_study 0.8021 1.2467 0.6830 0.9420
freq_cons_dum_5day 1.0313 0.9696 0.9421 1.1290
cond_oc_dum_2inact 1.0207 0.9797 0.9072 1.1484
cond_oc_dum_3unemp 0.9897 1.0104 0.8753 1.1190
susprindum_oh 1.0351 0.9661 0.7323 1.4631
susprindum_coc 1.0718 0.9330 0.7415 1.5494
susprindum_pbc 1.1163 0.8958 0.7755 1.6069
susprindum_mar 1.0555 0.9475 0.7082 1.5731
Concordance= 0.697 (se = 0.006 )
Likelihood ratio test= 817.7 on 20 df, p=<2e-16
Wald test = 852.1 on 20 df, p=<2e-16
Score (logrank) test = 848 on 20 df, p=<2e-16, Robust = 632.9 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).Code
# Call:
# coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
# log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
# policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
# susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar, data = datacox, cluster = id)
#
# n= 2342, number of events= 1597
# (163 observations deleted due to missingness)
#
# coef exp(coef) se(coef) robust se z Pr(>|z|)
# tr_outcome_rec.lag 2.07668 7.97793 0.18951 0.17050 12.180 < 2e-16 ***
# log_dias_treat_imp_sin_na.lag 0.48119 1.61800 0.04878 0.05862 8.208 2.24e-16 ***
# less_90d_tr1_rec.lag 0.53253 1.70324 0.15975 0.12591 4.230 2.34e-05 ***
# comp_bpsc_y3_severe_rec.lag 0.24739 1.28067 0.13099 0.12212 2.026 0.04280 *
# policonsumo2.lag -0.06660 0.93557 0.07911 0.07156 -0.931 0.35204
# edad_al_ing_1 0.10711 1.11305 0.01271 0.01175 9.116 < 2e-16 ***
# ano_nac_corr 0.11118 1.11759 0.01230 0.01146 9.705 < 2e-16 ***
# susinidum_coc_rec2 -0.19912 0.81945 0.16603 0.16462 -1.210 0.22642
# susinidum_oh -0.24978 0.77897 0.14191 0.12065 -2.070 0.03842 *
# susinidum_pbc -0.07015 0.93225 0.15381 0.13115 -0.535 0.59273
# susinidum_mar -0.14500 0.86502 0.14235 0.11908 -1.218 0.22334
# psycom_dum_with_rec2 -0.03390 0.96667 0.05027 0.04078 -0.831 0.40585
# psycom_dum_study -0.22053 0.80210 0.08298 0.08202 -2.689 0.00717 **
# freq_cons_dum_5day 0.03082 1.03130 0.05150 0.04617 0.668 0.50442
# cond_oc_dum_2inact 0.02049 1.02070 0.06522 0.06013 0.341 0.73326
# cond_oc_dum_3unemp -0.01038 0.98968 0.06734 0.06267 -0.166 0.86850
# susprindum_oh 0.03448 1.03508 0.18281 0.17658 0.195 0.84517
# susprindum_coc 0.06938 1.07184 0.18629 0.18800 0.369 0.71210
# susprindum_pbc 0.11002 1.11631 0.18066 0.18587 0.592 0.55390
# susprindum_mar 0.05398 1.05546 0.19901 0.20360 0.265 0.79092
# ---
# Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#
# exp(coef) exp(-coef) lower .95 upper .95
# tr_outcome_rec.lag 7.9779 0.1253 5.7117 11.1434
# log_dias_treat_imp_sin_na.lag 1.6180 0.6180 1.4424 1.8150
# less_90d_tr1_rec.lag 1.7032 0.5871 1.3308 2.1799
# comp_bpsc_y3_severe_rec.lag 1.2807 0.7808 1.0081 1.6270
# policonsumo2.lag 0.9356 1.0689 0.8131 1.0764
# edad_al_ing_1 1.1131 0.8984 1.0877 1.1390
# ano_nac_corr 1.1176 0.8948 1.0928 1.1430
# susinidum_coc_rec2 0.8194 1.2203 0.5935 1.1315
# susinidum_oh 0.7790 1.2837 0.6149 0.9868
# susinidum_pbc 0.9323 1.0727 0.7209 1.2055
# susinidum_mar 0.8650 1.1560 0.6850 1.0924
# psycom_dum_with_rec2 0.9667 1.0345 0.8924 1.0471
# psycom_dum_study 0.8021 1.2467 0.6830 0.9420
# freq_cons_dum_5day 1.0313 0.9696 0.9421 1.1290
# cond_oc_dum_2inact 1.0207 0.9797 0.9072 1.1484
# cond_oc_dum_3unemp 0.9897 1.0104 0.8753 1.1190
# susprindum_oh 1.0351 0.9661 0.7323 1.4631
# susprindum_coc 1.0718 0.9330 0.7415 1.5494
# susprindum_pbc 1.1163 0.8958 0.7755 1.6069
# susprindum_mar 1.0555 0.9475 0.7082 1.5731
#
# Concordance= 0.697 (se = 0.006 )
# Likelihood ratio test= 817.7 on 20 df, p=<2e-16
# Wald test = 852.1 on 20 df, p=<2e-16
# Score (logrank) test = 848 on 20 df, p=<2e-16, Robust = 632.9 p=<2e-16
#
# (Note: the likelihood ratio and score tests assume independence of
# observations within a cluster, the Wald and robust score tests do not).
summary(iiw_model_after_ph_wo_ia_alt$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.1065 0.2508 0.7104 1.3025 1.8834 27.4746 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.1065 0.2508 0.7104 1.3025 1.8834 27.4746
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("WO residential")
iiw_model_after_ph_wo_res<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1_rec.lag+
comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod== levels_tipo_de_plan_2_mod[[5]]) %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= subset(maxfu_restr_df, subset= maxfu.id %in% subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod==levels_tipo_de_plan_2_mod[[5]])$id) %>% as.data.frame(),
# formulanull = Surv(time.lag,time,event)~ edad_al_ing_1 +
# ano_nac_corr,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc_rec2", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with_rec2", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome_rec", "comp_bpsc_y3_severe_rec", "less_90d_tr1_rec", "policonsumo2"),
lagfirst= c(2.95082,4.499811/2,0,0,0,0), #90/30.5 4.499811 es 90 días
first= T
)
summary(iiw_model_after_ph_wo_res$m)Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 3189, number of events= 2153
(381 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome_rec.lag -0.121612 0.885492 0.165763 0.129366 -0.940
log_dias_treat_imp_sin_na.lag -0.130201 0.877919 0.014462 0.014401 -9.041
less_90d_tr1_rec.lag 0.090970 1.095236 0.149749 0.121156 0.751
comp_bpsc_y3_severe_rec.lag 0.404439 1.498462 0.142265 0.111507 3.627
policonsumo2.lag -0.669071 0.512184 0.052846 0.047637 -14.045
edad_al_ing_1 0.033148 1.033703 0.010263 0.006975 4.752
ano_nac_corr 0.039915 1.040723 0.009991 0.006723 5.937
susinidum_coc_rec2 0.423752 1.527683 0.156368 0.113984 3.718
susinidum_oh 0.340455 1.405586 0.131706 0.085631 3.976
susinidum_pbc 0.248609 1.282241 0.138130 0.094813 2.622
susinidum_mar 0.339920 1.404836 0.132224 0.085995 3.953
psycom_dum_with_rec2 0.080161 1.083462 0.047830 0.034178 2.345
psycom_dum_study -0.359226 0.698216 0.069917 0.058469 -6.144
freq_cons_dum_5day -0.014932 0.985179 0.050738 0.034074 -0.438
cond_oc_dum_2inact 0.017569 1.017725 0.081296 0.053721 0.327
cond_oc_dum_3unemp 0.089596 1.093733 0.081392 0.054694 1.638
susprindum_oh -0.081599 0.921642 0.147882 0.078177 -1.044
susprindum_coc 0.046968 1.048088 0.149001 0.085066 0.552
susprindum_pbc -0.045903 0.955134 0.139802 0.074248 -0.618
susprindum_mar 0.120782 1.128378 0.210478 0.124728 0.968
Pr(>|z|)
tr_outcome_rec.lag 0.347187
log_dias_treat_imp_sin_na.lag < 2e-16 ***
less_90d_tr1_rec.lag 0.452744
comp_bpsc_y3_severe_rec.lag 0.000287 ***
policonsumo2.lag < 2e-16 ***
edad_al_ing_1 2.01e-06 ***
ano_nac_corr 2.89e-09 ***
susinidum_coc_rec2 0.000201 ***
susinidum_oh 7.01e-05 ***
susinidum_pbc 0.008739 **
susinidum_mar 7.72e-05 ***
psycom_dum_with_rec2 0.019005 *
psycom_dum_study 8.05e-10 ***
freq_cons_dum_5day 0.661225
cond_oc_dum_2inact 0.743632
cond_oc_dum_3unemp 0.101394
susprindum_oh 0.296592
susprindum_coc 0.580858
susprindum_pbc 0.536413
susprindum_mar 0.332866
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome_rec.lag 0.8855 1.1293 0.6872 1.1410
log_dias_treat_imp_sin_na.lag 0.8779 1.1391 0.8535 0.9031
less_90d_tr1_rec.lag 1.0952 0.9130 0.8637 1.3888
comp_bpsc_y3_severe_rec.lag 1.4985 0.6674 1.2043 1.8645
policonsumo2.lag 0.5122 1.9524 0.4665 0.5623
edad_al_ing_1 1.0337 0.9674 1.0197 1.0479
ano_nac_corr 1.0407 0.9609 1.0271 1.0545
susinidum_coc_rec2 1.5277 0.6546 1.2218 1.9101
susinidum_oh 1.4056 0.7114 1.1884 1.6624
susinidum_pbc 1.2822 0.7799 1.0648 1.5441
susinidum_mar 1.4048 0.7118 1.1869 1.6627
psycom_dum_with_rec2 1.0835 0.9230 1.0133 1.1585
psycom_dum_study 0.6982 1.4322 0.6226 0.7830
freq_cons_dum_5day 0.9852 1.0150 0.9215 1.0532
cond_oc_dum_2inact 1.0177 0.9826 0.9160 1.1307
cond_oc_dum_3unemp 1.0937 0.9143 0.9826 1.2175
susprindum_oh 0.9216 1.0850 0.7907 1.0743
susprindum_coc 1.0481 0.9541 0.8871 1.2382
susprindum_pbc 0.9551 1.0470 0.8258 1.1048
susprindum_mar 1.1284 0.8862 0.8837 1.4409
Concordance= 0.612 (se = 0.005 )
Likelihood ratio test= 365.4 on 20 df, p=<2e-16
Wald test = 543 on 20 df, p=<2e-16
Score (logrank) test = 379.9 on 20 df, p=<2e-16, Robust = 480.8 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).Code
# Call:
# coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
# log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
# policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
# susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar, data = datacox, cluster = id)
#
# n= 3189, number of events= 2153
# (381 observations deleted due to missingness)
#
# coef exp(coef) se(coef) robust se z Pr(>|z|)
# tr_outcome_rec.lag -0.126675 0.881020 0.163284 0.124402 -1.018 0.308550
# log_dias_treat_imp_sin_na.lag 0.031652 1.032159 0.022850 0.021963 1.441 0.149546
# less_90d_tr1_rec.lag 0.363190 1.437908 0.157358 0.121227 2.996 0.002736 **
# comp_bpsc_y3_severe_rec.lag 0.349023 1.417682 0.139042 0.107539 3.246 0.001172 **
# policonsumo2.lag -0.778676 0.459014 0.050648 0.044979 -17.312 < 2e-16 ***
# edad_al_ing_1 0.023880 1.024168 0.010199 0.006895 3.463 0.000534 ***
# ano_nac_corr 0.030909 1.031392 0.009928 0.006625 4.666 3.07e-06 ***
# susinidum_coc_rec2 0.353985 1.424734 0.156986 0.101414 3.490 0.000482 ***
# susinidum_oh 0.291847 1.338899 0.132369 0.078022 3.741 0.000184 ***
# susinidum_pbc 0.186272 1.204750 0.138703 0.088234 2.111 0.034763 *
# susinidum_mar 0.294218 1.342076 0.132819 0.078394 3.753 0.000175 ***
# psycom_dum_with_rec2 0.075649 1.078584 0.048007 0.033882 2.233 0.025566 *
# psycom_dum_study -0.341685 0.710572 0.070190 0.057146 -5.979 2.24e-09 ***
# freq_cons_dum_5day -0.020624 0.979587 0.050671 0.033925 -0.608 0.543239
# cond_oc_dum_2inact 0.011996 1.012068 0.081341 0.053444 0.224 0.822402
# cond_oc_dum_3unemp 0.076550 1.079556 0.081384 0.054527 1.404 0.160354
# susprindum_oh -0.077120 0.925778 0.148321 0.077766 -0.992 0.321346
# susprindum_coc 0.072353 1.075035 0.149194 0.083542 0.866 0.386453
# susprindum_pbc -0.011860 0.988210 0.140132 0.073586 -0.161 0.871956
# susprindum_mar 0.154065 1.166566 0.210712 0.124652 1.236 0.216472
# ---
# Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#
# exp(coef) exp(-coef) lower .95 upper .95
# tr_outcome_rec.lag 0.8810 1.1350 0.6904 1.1243
# log_dias_treat_imp_sin_na.lag 1.0322 0.9688 0.9887 1.0776
# less_90d_tr1_rec.lag 1.4379 0.6955 1.1338 1.8236
# comp_bpsc_y3_severe_rec.lag 1.4177 0.7054 1.1483 1.7503
# policonsumo2.lag 0.4590 2.1786 0.4203 0.5013
# edad_al_ing_1 1.0242 0.9764 1.0104 1.0381
# ano_nac_corr 1.0314 0.9696 1.0181 1.0449
# susinidum_coc_rec2 1.4247 0.7019 1.1679 1.7380
# susinidum_oh 1.3389 0.7469 1.1490 1.5601
# susinidum_pbc 1.2047 0.8300 1.0134 1.4322
# susinidum_mar 1.3421 0.7451 1.1509 1.5650
# psycom_dum_with_rec2 1.0786 0.9271 1.0093 1.1526
# psycom_dum_study 0.7106 1.4073 0.6353 0.7948
# freq_cons_dum_5day 0.9796 1.0208 0.9166 1.0469
# cond_oc_dum_2inact 1.0121 0.9881 0.9114 1.1238
# cond_oc_dum_3unemp 1.0796 0.9263 0.9701 1.2013
# susprindum_oh 0.9258 1.0802 0.7949 1.0782
# susprindum_coc 1.0750 0.9302 0.9127 1.2663
# susprindum_pbc 0.9882 1.0119 0.8555 1.1415
# susprindum_mar 1.1666 0.8572 0.9137 1.4894
#
# Concordance= 0.612 (se = 0.005 )
# Likelihood ratio test= 298.2 on 20 df, p=<2e-16
# Wald test = 480.2 on 20 df, p=<2e-16
# Score (logrank) test = 312.9 on 20 df, p=<2e-16, Robust = 428.9 p=<2e-16
#
# (Note: the likelihood ratio and score tests assume independence of
# observations within a cluster, the Wald and robust score tests do not).
summary(iiw_model_after_ph_wo_res$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.2809 0.6732 1.0225 1.1916 1.6473 3.6104 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.4578 0.8652 1.2333 1.3288 1.6967 4.6168
iiw_model_after_ph_wo_res_alt<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1_rec.lag+
comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc_rec2 +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with_rec2 +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod== levels_tipo_de_plan_2_mod[[5]]) %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= subset(maxfu_restr_df, subset= maxfu.id %in% subset(data_mine_miss_restr_proc2, tipo_de_plan_2_mod==levels_tipo_de_plan_2_mod[[5]])$id) %>% as.data.frame(),
# formulanull = Surv(time.lag,time,event)~ edad_al_ing_1 +
# ano_nac_corr,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc_rec2", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with_rec2", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome_rec", "comp_bpsc_y3_severe_rec", "less_90d_tr1_rec", "policonsumo2"),
lagfirst= c(2.95082,4.499811,1,1,1,1), #90/30.5 4.499811 es 90 días
first= T
)
summary(iiw_model_after_ph_wo_res_alt$m)Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 3189, number of events= 2153
(381 observations deleted due to missingness)
coef exp(coef) se(coef) robust se
tr_outcome_rec.lag 0.7205801 2.0556254 0.1852413 0.1688667
log_dias_treat_imp_sin_na.lag 0.2567171 1.2926793 0.0273630 0.0340385
less_90d_tr1_rec.lag 0.6130288 1.8460141 0.1493355 0.1272692
comp_bpsc_y3_severe_rec.lag 0.8473593 2.3334766 0.1535169 0.1444312
policonsumo2.lag 0.0184737 1.0186454 0.0734914 0.0667181
edad_al_ing_1 0.0575836 1.0592739 0.0104369 0.0089922
ano_nac_corr 0.0650981 1.0672637 0.0101748 0.0087168
susinidum_coc_rec2 0.3116140 1.3656274 0.1602666 0.1367453
susinidum_oh 0.1895267 1.2086774 0.1337491 0.0914487
susinidum_pbc 0.1983614 1.2194030 0.1404151 0.1021221
susinidum_mar 0.1989101 1.2200723 0.1343948 0.0922898
psycom_dum_with_rec2 -0.0087068 0.9913310 0.0483988 0.0404580
psycom_dum_study -0.3746894 0.6875028 0.0680287 0.0652967
freq_cons_dum_5day 0.0004056 1.0004056 0.0506595 0.0444514
cond_oc_dum_2inact 0.0813419 1.0847417 0.0809746 0.0759279
cond_oc_dum_3unemp 0.1300692 1.1389072 0.0814469 0.0765984
susprindum_oh -0.1079311 0.8976894 0.1483977 0.1155415
susprindum_coc 0.0027078 1.0027114 0.1498413 0.1207600
susprindum_pbc -0.0637471 0.9382423 0.1407408 0.1099733
susprindum_mar -0.0838286 0.9195888 0.2112160 0.1592995
z Pr(>|z|)
tr_outcome_rec.lag 4.267 1.98e-05 ***
log_dias_treat_imp_sin_na.lag 7.542 4.63e-14 ***
less_90d_tr1_rec.lag 4.817 1.46e-06 ***
comp_bpsc_y3_severe_rec.lag 5.867 4.44e-09 ***
policonsumo2.lag 0.277 0.7819
edad_al_ing_1 6.404 1.52e-10 ***
ano_nac_corr 7.468 8.14e-14 ***
susinidum_coc_rec2 2.279 0.0227 *
susinidum_oh 2.072 0.0382 *
susinidum_pbc 1.942 0.0521 .
susinidum_mar 2.155 0.0311 *
psycom_dum_with_rec2 -0.215 0.8296
psycom_dum_study -5.738 9.57e-09 ***
freq_cons_dum_5day 0.009 0.9927
cond_oc_dum_2inact 1.071 0.2840
cond_oc_dum_3unemp 1.698 0.0895 .
susprindum_oh -0.934 0.3502
susprindum_coc 0.022 0.9821
susprindum_pbc -0.580 0.5621
susprindum_mar -0.526 0.5987
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome_rec.lag 2.0556 0.4865 1.4764 2.8621
log_dias_treat_imp_sin_na.lag 1.2927 0.7736 1.2093 1.3819
less_90d_tr1_rec.lag 1.8460 0.5417 1.4385 2.3690
comp_bpsc_y3_severe_rec.lag 2.3335 0.4285 1.7582 3.0970
policonsumo2.lag 1.0186 0.9817 0.8938 1.1610
edad_al_ing_1 1.0593 0.9440 1.0408 1.0781
ano_nac_corr 1.0673 0.9370 1.0492 1.0857
susinidum_coc_rec2 1.3656 0.7323 1.0446 1.7854
susinidum_oh 1.2087 0.8274 1.0103 1.4459
susinidum_pbc 1.2194 0.8201 0.9982 1.4896
susinidum_mar 1.2201 0.8196 1.0182 1.4620
psycom_dum_with_rec2 0.9913 1.0087 0.9158 1.0731
psycom_dum_study 0.6875 1.4545 0.6049 0.7814
freq_cons_dum_5day 1.0004 0.9996 0.9169 1.0915
cond_oc_dum_2inact 1.0847 0.9219 0.9348 1.2588
cond_oc_dum_3unemp 1.1389 0.8780 0.9801 1.3234
susprindum_oh 0.8977 1.1140 0.7158 1.1258
susprindum_coc 1.0027 0.9973 0.7914 1.2705
susprindum_pbc 0.9382 1.0658 0.7563 1.1639
susprindum_mar 0.9196 1.0874 0.6730 1.2566
Concordance= 0.666 (se = 0.005 )
Likelihood ratio test= 814.4 on 20 df, p=<2e-16
Wald test = 982.1 on 20 df, p=<2e-16
Score (logrank) test = 834.4 on 20 df, p=<2e-16, Robust = 780.6 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).Code
# Call:
# coxph(formula = Surv(time.lag, time, event) ~ tr_outcome_rec.lag +
# log_dias_treat_imp_sin_na.lag + less_90d_tr1_rec.lag + comp_bpsc_y3_severe_rec.lag +
# policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_coc_rec2 +
# susinidum_oh + susinidum_pbc + susinidum_mar + psycom_dum_with_rec2 +
# psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar, data = datacox, cluster = id)
#
# n= 3189, number of events= 2153
# (381 observations deleted due to missingness)
#
# coef exp(coef) se(coef) robust se z Pr(>|z|)
# tr_outcome_rec.lag 0.7205801 2.0556254 0.1852413 0.1688667 4.267 1.98e-05 ***
# log_dias_treat_imp_sin_na.lag 0.2567171 1.2926793 0.0273630 0.0340385 7.542 4.63e-14 ***
# less_90d_tr1_rec.lag 0.6130288 1.8460141 0.1493355 0.1272692 4.817 1.46e-06 ***
# comp_bpsc_y3_severe_rec.lag 0.8473593 2.3334766 0.1535169 0.1444312 5.867 4.44e-09 ***
# policonsumo2.lag 0.0184737 1.0186454 0.0734914 0.0667181 0.277 0.7819
# edad_al_ing_1 0.0575836 1.0592739 0.0104369 0.0089922 6.404 1.52e-10 ***
# ano_nac_corr 0.0650981 1.0672637 0.0101748 0.0087168 7.468 8.14e-14 ***
# susinidum_coc_rec2 0.3116140 1.3656274 0.1602666 0.1367453 2.279 0.0227 *
# susinidum_oh 0.1895267 1.2086774 0.1337491 0.0914487 2.072 0.0382 *
# susinidum_pbc 0.1983614 1.2194030 0.1404151 0.1021221 1.942 0.0521 .
# susinidum_mar 0.1989101 1.2200723 0.1343948 0.0922898 2.155 0.0311 *
# psycom_dum_with_rec2 -0.0087068 0.9913310 0.0483988 0.0404580 -0.215 0.8296
# psycom_dum_study -0.3746894 0.6875028 0.0680287 0.0652967 -5.738 9.57e-09 ***
# freq_cons_dum_5day 0.0004056 1.0004056 0.0506595 0.0444514 0.009 0.9927
# cond_oc_dum_2inact 0.0813419 1.0847417 0.0809746 0.0759279 1.071 0.2840
# cond_oc_dum_3unemp 0.1300692 1.1389072 0.0814469 0.0765984 1.698 0.0895 .
# susprindum_oh -0.1079311 0.8976894 0.1483977 0.1155415 -0.934 0.3502
# susprindum_coc 0.0027078 1.0027114 0.1498413 0.1207600 0.022 0.9821
# susprindum_pbc -0.0637471 0.9382423 0.1407408 0.1099733 -0.580 0.5621
# susprindum_mar -0.0838286 0.9195888 0.2112160 0.1592995 -0.526 0.5987
# ---
# Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#
# exp(coef) exp(-coef) lower .95 upper .95
# tr_outcome_rec.lag 2.0556 0.4865 1.4764 2.8621
# log_dias_treat_imp_sin_na.lag 1.2927 0.7736 1.2093 1.3819
# less_90d_tr1_rec.lag 1.8460 0.5417 1.4385 2.3690
# comp_bpsc_y3_severe_rec.lag 2.3335 0.4285 1.7582 3.0970
# policonsumo2.lag 1.0186 0.9817 0.8938 1.1610
# edad_al_ing_1 1.0593 0.9440 1.0408 1.0781
# ano_nac_corr 1.0673 0.9370 1.0492 1.0857
# susinidum_coc_rec2 1.3656 0.7323 1.0446 1.7854
# susinidum_oh 1.2087 0.8274 1.0103 1.4459
# susinidum_pbc 1.2194 0.8201 0.9982 1.4896
# susinidum_mar 1.2201 0.8196 1.0182 1.4620
# psycom_dum_with_rec2 0.9913 1.0087 0.9158 1.0731
# psycom_dum_study 0.6875 1.4545 0.6049 0.7814
# freq_cons_dum_5day 1.0004 0.9996 0.9169 1.0915
# cond_oc_dum_2inact 1.0847 0.9219 0.9348 1.2588
# cond_oc_dum_3unemp 1.1389 0.8780 0.9801 1.3234
# susprindum_oh 0.8977 1.1140 0.7158 1.1258
# susprindum_coc 1.0027 0.9973 0.7914 1.2705
# susprindum_pbc 0.9382 1.0658 0.7563 1.1639
# susprindum_mar 0.9196 1.0874 0.6730 1.2566
#
# Concordance= 0.666 (se = 0.005 )
# Likelihood ratio test= 814.4 on 20 df, p=<2e-16
# Wald test = 982.1 on 20 df, p=<2e-16
# Score (logrank) test = 834.4 on 20 df, p=<2e-16, Robust = 780.6 p=<2e-16
#
# (Note: the likelihood ratio and score tests assume independence of
# observations within a cluster, the Wald and robust score tests do not).
summary(iiw_model_after_ph_wo_res_alt$iiw.weight) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.1235 0.2374 0.5581 0.8884 1.2695 52.2809 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.1235 0.2374 0.5581 0.8884 1.2695 52.2809 Code
data_mine_miss_restr_proc2_iiw_after_ph <- NULL # Initialize empty result
df_models_after_ph<-
cbind.data.frame(tipo_de_plan_2= attr(table(data_mine_miss_restr_proc2$tipo_de_plan_2_mod),"names"),
models_after_ph= c("iiw_model_after_ph_ba",
"iiw_model_after_ph_gp_ia",
"iiw_model_after_ph_gp_res",
"iiw_model_after_ph_wo_ia",
"iiw_model_after_ph_wo_res"))
for (i in 1:nrow(df_models_after_ph)) {
subset_data0 <- subset(dplyr::mutate(data_mine_miss_restr_proc2, rn=dplyr::row_number()), subset= tipo_de_plan_2_mod== df_models_after_ph[i,1])
subset_data02<- cbind.data.frame(subset_data0, iiw_after_ph= get(df_models_after_ph[i,2])$iiw.weight)
data_mine_miss_restr_proc2_iiw_after_ph <- rbind.data.frame(data_mine_miss_restr_proc2_iiw_after_ph, subset_data02)
}
#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_
data_mine_miss_restr_proc2_iiw_after_ph_alt <- NULL # Initialize empty result
df_models_after_ph_alt<-
cbind.data.frame(tipo_de_plan_2= attr(table(data_mine_miss_restr_proc2$tipo_de_plan_2_mod),"names"),
models_after_ph= paste0(c("iiw_model_after_ph_ba",
"iiw_model_after_ph_gp_ia",
"iiw_model_after_ph_gp_res",
"iiw_model_after_ph_wo_ia",
"iiw_model_after_ph_wo_res"),"_alt"))
for (i in 1:nrow(df_models_after_ph_alt)) {
subset_data00 <- subset(dplyr::mutate(data_mine_miss_restr_proc2, rn=dplyr::row_number()), subset= tipo_de_plan_2_mod== df_models_after_ph_alt[i,1])
subset_data002<- cbind.data.frame(subset_data00, iiw_after_ph= get(df_models_after_ph_alt[i,2])$iiw.weight)
data_mine_miss_restr_proc2_iiw_after_ph_alt <- rbind.data.frame(data_mine_miss_restr_proc2_iiw_after_ph_alt, subset_data002)
}
summary(data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.09172 0.66527 1.08442 1.33483 1.90360 5.42701 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.4578 0.8791 1.1600 1.3930 1.8724 6.7300
summary(data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.0643 0.1688 0.5455 1.1908 1.3579 1294.5167 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.0643 0.1688 0.5455 1.1908 1.3579 1294.5167
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Stabilize weights")
data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph_st<-data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph
data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph_st[data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph>quantile(data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph,0.975)] <- quantile(data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph,0.975)
data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph_st[data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph<quantile(data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph,0.025)] <- quantile(data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph,0.025)
invisible("stabilize alternative weights")
data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph_st<-data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph
data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph_st[data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph>quantile(data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph,0.975)] <- quantile(data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph,0.975)
data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph_st[data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph<quantile(data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph,0.025)] <- quantile(data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph,0.025)
summary(data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph_st) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.4682 0.6653 1.0844 1.3266 1.9036 3.0250 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.6912 0.8791 1.1600 1.3873 1.8724 2.5741
summary(data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph_st) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.09766 0.16882 0.54545 0.88019 1.35785 3.56260 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.09766 0.16882 0.54545 0.88019 1.35785 3.56260
rbind.data.frame(
cbind.data.frame(model= "Primary (after PH, lag=0)", t(matrix(summary(data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph_st)))),
cbind.data.frame(model= "Alternative (after PH, lag=1)", t(matrix(summary(data_mine_miss_restr_proc2_iiw_after_ph_alt$iiw_after_ph_st)))))%>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/iiw_after_ph_240426.csv"), dec=",", sep="\t")
knitr::kable(., size=10, format="markdown", caption="Weights (after correction for PHs)", col.names= c("Weight",attr(summary(data_mine_miss_restr_proc2_iiw_after_ph$iiw_after_ph_st),"names")) )
}| Weight | Min. | 1st Qu. | Median | Mean | 3rd Qu. | Max. |
|---|---|---|---|---|---|---|
| Primary (after PH, lag=0) | 0.4681976 | 0.6652737 | 1.084424 | 1.3266201 | 1.903597 | 3.025044 |
| Alternative (after PH, lag=1) | 0.0976600 | 0.1688223 | 0.545454 | 0.8801874 | 1.357854 | 3.562599 |
GEE
Code
#The id= option is where we specify the clusters within which we have repeated observations that may be correlated, and tehe scale.fix=T option is to avoid R's default approach of introducing an overdispersion scale parameter.
#However, caution is advised when employing QIC and its use should not be routine, see Wang et al. (2015). growing use of quasi-likelihood-based information criteria for longitudinal data to select a working correlation structure in a generalized estimating equation framework. [https://onlinelibrary.wiley.com/doi/full/10.1002/sta4.95]
#poisson(link = "log"),
#https://www.thelancet.com/cms/10.1016/S2468-2667(22)00042-1/attachment/eb69ffc4-cb6c-4a47-9524-94e1ba9302b2/mmc1.pdf
#https://www.thelancet.com/cms/10.1016/S2468-2667(22)00201-8/attachment/05a17ae1-1e31-4ffd-a727-78ec915dc03b/mmc1.pdf
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Must be independence structure")
m <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_restr_proc2_iiw_after_ph, family=binomial, corstr="independence")
summary(m)
Call:
geeglm(formula = tr_outcome ~ policonsumo2 + edad_al_ing_1 +
ano_nac_corr + susinidum_oh + susinidum_coc + susinidum_pbc +
susinidum_mar + psycom_dum_study + psycom_dum_with + freq_cons_dum_5day +
cond_oc_dum_2inact + cond_oc_dum_3unemp + susprindum_coc +
susprindum_pbc + susprindum_mar, family = binomial, data = data_mine_miss_restr_proc2_iiw_after_ph,
id = id, corstr = "independence")
Coefficients:
Estimate Std.err Wald Pr(>|W|)
(Intercept) -43.338627 13.168777 10.831 0.000998 ***
policonsumo2 0.057600 0.035967 2.565 0.109271
edad_al_ing_1 -0.006693 0.006676 1.005 0.316079
ano_nac_corr 0.022507 0.006543 11.831 0.000583 ***
susinidum_oh -0.073882 0.094343 0.613 0.433556
susinidum_coc -0.015069 0.118072 0.016 0.898446
susinidum_pbc 0.145421 0.107807 1.820 0.177366
susinidum_mar 0.047494 0.096584 0.242 0.622904
psycom_dum_study 0.863462 0.046317 347.538 < 2e-16 ***
psycom_dum_with 0.060458 0.032631 3.433 0.063916 .
freq_cons_dum_5day -0.029792 0.030889 0.930 0.334802
cond_oc_dum_2inact -0.258428 0.042900 36.288 1.70e-09 ***
cond_oc_dum_3unemp -0.192816 0.034495 31.245 2.27e-08 ***
susprindum_coc 0.029585 0.048404 0.374 0.541054
susprindum_pbc 0.237070 0.040895 33.606 6.75e-09 ***
susprindum_mar 0.062763 0.074961 0.701 0.402438
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation structure = independence
Estimated Scale Parameters:
Estimate Std.err
(Intercept) 1.001 0.01624
Number of clusters: 13317 Maximum cluster size: 10 Code
m_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_restr_proc2_iiw_after_ph, family=binomial, corstr="independence")
rbind.data.frame(cbind.data.frame(model="simplest", broom::tidy(m, exponentiate = T, conf.int=T)),cbind.data.frame(model="full", broom::tidy(m_full, exponentiate = T, conf.int=T))) %>% dplyr::mutate_at(c("estimate","std.error","conf.low","conf.high"), ~sprintf("%1.2f",.))%>% dplyr::mutate_at(c("p.value"), ~sprintf("%1.4f",.)) %>% dplyr::select(-statistic) %>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/gee_240502.csv"), dec=",", sep="\t")
knitr::kable(dplyr::filter(.,term=="policonsumo2"), size=10, format="markdown", caption="GEE Models (after correction for PHs)")
}| model | term | estimate | std.error | p.value | conf.low | conf.high |
|---|---|---|---|---|---|---|
| simplest | policonsumo2 | 1.06 | 0.04 | 0.1093 | 0.99 | 1.14 |
| full | policonsumo2 | 1.10 | 0.04 | 0.0079 | 1.03 | 1.18 |
Code
data_mine_miss_restr_proc2_iiw_after_ph$edad_al_ing_1_st = data_mine_miss_restr_proc2_iiw_after_ph$edad_al_ing_1 - mean(data_mine_miss_restr_proc2_iiw_after_ph$edad_al_ing_1)
# can use to fit a weighted GEE
mw <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_restr_proc2_iiw_after_ph, family= binomial, weights=iiw_after_ph_st, corstr="independence")Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Code
#In eval(family$initialize) : non-integer #successes in a binomial glm!
#summary(mw)
mwpois <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_restr_proc2_iiw_after_ph, family= poisson(link = "log"), weights=iiw_after_ph_st, corstr="independence")
summary(mwpois)
Call:
geeglm(formula = tr_outcome ~ policonsumo2 + edad_al_ing_1 +
ano_nac_corr + susinidum_oh + susinidum_coc + susinidum_pbc +
susinidum_mar + psycom_dum_study + psycom_dum_with + freq_cons_dum_5day +
cond_oc_dum_2inact + cond_oc_dum_3unemp + susprindum_coc +
susprindum_pbc + susprindum_mar, family = poisson(link = "log"),
data = data_mine_miss_restr_proc2_iiw_after_ph, weights = iiw_after_ph_st,
id = id, corstr = "independence")
Coefficients:
Estimate Std.err Wald Pr(>|W|)
(Intercept) -6.48845 3.56281 3.32 0.0686 .
policonsumo2 0.02903 0.01098 6.99 0.0082 **
edad_al_ing_1 -0.00520 0.00182 8.13 0.0044 **
ano_nac_corr 0.00318 0.00177 3.23 0.0724 .
susinidum_oh -0.01208 0.02868 0.18 0.6736
susinidum_coc 0.00578 0.03446 0.03 0.8668
susinidum_pbc 0.03984 0.03091 1.66 0.1975
susinidum_mar 0.01794 0.02878 0.39 0.5331
psycom_dum_study 0.13012 0.01024 161.40 < 2e-16 ***
psycom_dum_with 0.00609 0.01002 0.37 0.5436
freq_cons_dum_5day 0.00429 0.00860 0.25 0.6179
cond_oc_dum_2inact -0.05450 0.01246 19.14 1.2e-05 ***
cond_oc_dum_3unemp -0.03855 0.00943 16.71 4.4e-05 ***
susprindum_coc 0.01730 0.01487 1.35 0.2447
susprindum_pbc 0.05929 0.01247 22.60 2.0e-06 ***
susprindum_mar 0.01361 0.02142 0.40 0.5252
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation structure = independence
Estimated Scale Parameters:
Estimate Std.err
(Intercept) 0.256 0.00339
Number of clusters: 13317 Maximum cluster size: 10 Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
mw_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_restr_proc2_iiw_after_ph, family=binomial, weights=iiw_after_ph_st, corstr="independence") Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Code
#In eval(family$initialize) : non-integer #successes in a binomial glm!
mwpois_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_restr_proc2_iiw_after_ph, family= poisson(link = "log"), weights=iiw_after_ph_st, corstr="independence")
invisible("Model fit")
#broom::glance(mwpois_full)
#QIC(mw)
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
rbind.data.frame(cbind.data.frame(model="simplest, binomial", broom::tidy(mw, exponentiate = T, conf.int=T)),cbind.data.frame(model="full, binomial", broom::tidy(mw_full, exponentiate = T, conf.int=T)),cbind.data.frame(model="simplest, poisson", broom::tidy(mwpois, exponentiate = T, conf.int=T)),cbind.data.frame(model="full, poisson", broom::tidy(mwpois_full, exponentiate = T, conf.int=T))) %>% dplyr::mutate_at(c("estimate","std.error","conf.low","conf.high"), ~sprintf("%1.2f",.))%>% dplyr::mutate_at(c("p.value"), ~sprintf("%1.4f",.)) %>% dplyr::select(-statistic) %>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/geew_240502.csv"), dec=",", sep="\t")
knitr::kable(dplyr::filter(.,term=="policonsumo2"), size=10, format="markdown", caption="GEE Models, weighted (after correction for PHs)")
}| model | term | estimate | std.error | p.value | conf.low | conf.high |
|---|---|---|---|---|---|---|
| simplest, binomial | policonsumo2 | 1.10 | 0.04 | 0.0238 | 1.01 | 1.19 |
| full, binomial | policonsumo2 | 1.13 | 0.04 | 0.0032 | 1.04 | 1.22 |
| simplest, poisson | policonsumo2 | 1.03 | 0.01 | 0.0082 | 1.01 | 1.05 |
| full, poisson | policonsumo2 | 1.04 | 0.01 | 0.0013 | 1.01 | 1.06 |
Code
# can use to fit a weighted GEE
mw_alt <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_restr_proc2_iiw_after_ph_alt, family= binomial, weights=iiw_after_ph_st, corstr="independence")Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Code
#In eval(family$initialize) : non-integer #successes in a binomial glm!
#summary(mw)
mwpois_alt <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_restr_proc2_iiw_after_ph_alt, family= poisson(link = "log"), weights=iiw_after_ph_st, corstr="independence")
summary(mwpois)
Call:
geeglm(formula = tr_outcome ~ policonsumo2 + edad_al_ing_1 +
ano_nac_corr + susinidum_oh + susinidum_coc + susinidum_pbc +
susinidum_mar + psycom_dum_study + psycom_dum_with + freq_cons_dum_5day +
cond_oc_dum_2inact + cond_oc_dum_3unemp + susprindum_coc +
susprindum_pbc + susprindum_mar, family = poisson(link = "log"),
data = data_mine_miss_restr_proc2_iiw_after_ph, weights = iiw_after_ph_st,
id = id, corstr = "independence")
Coefficients:
Estimate Std.err Wald Pr(>|W|)
(Intercept) -6.48845 3.56281 3.32 0.0686 .
policonsumo2 0.02903 0.01098 6.99 0.0082 **
edad_al_ing_1 -0.00520 0.00182 8.13 0.0044 **
ano_nac_corr 0.00318 0.00177 3.23 0.0724 .
susinidum_oh -0.01208 0.02868 0.18 0.6736
susinidum_coc 0.00578 0.03446 0.03 0.8668
susinidum_pbc 0.03984 0.03091 1.66 0.1975
susinidum_mar 0.01794 0.02878 0.39 0.5331
psycom_dum_study 0.13012 0.01024 161.40 < 2e-16 ***
psycom_dum_with 0.00609 0.01002 0.37 0.5436
freq_cons_dum_5day 0.00429 0.00860 0.25 0.6179
cond_oc_dum_2inact -0.05450 0.01246 19.14 1.2e-05 ***
cond_oc_dum_3unemp -0.03855 0.00943 16.71 4.4e-05 ***
susprindum_coc 0.01730 0.01487 1.35 0.2447
susprindum_pbc 0.05929 0.01247 22.60 2.0e-06 ***
susprindum_mar 0.01361 0.02142 0.40 0.5252
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation structure = independence
Estimated Scale Parameters:
Estimate Std.err
(Intercept) 0.256 0.00339
Number of clusters: 13317 Maximum cluster size: 10 Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
mw_alt_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_restr_proc2_iiw_after_ph_alt, family=binomial, weights=iiw_after_ph_st, corstr="independence") Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Code
#In eval(family$initialize) : non-integer #successes in a binomial glm!
mwpois_alt_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_restr_proc2_iiw_after_ph_alt, family= poisson(link = "log"), weights=iiw_after_ph_st, corstr="independence")
invisible("Model fit")
#broom::glance(mwpois_full)
#QIC(mw)
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
rbind.data.frame(cbind.data.frame(model="simplest, binomial", broom::tidy(mw_alt, exponentiate = T, conf.int=T)),cbind.data.frame(model="full, binomial", broom::tidy(mw_alt_full, exponentiate = T, conf.int=T)),cbind.data.frame(model="simplest, poisson", broom::tidy(mwpois_alt, exponentiate = T, conf.int=T)),cbind.data.frame(model="full, poisson", broom::tidy(mwpois_alt_full, exponentiate = T, conf.int=T))) %>% dplyr::mutate_at(c("estimate","std.error","conf.low","conf.high"), ~sprintf("%1.2f",.))%>% dplyr::mutate_at(c("p.value"), ~sprintf("%1.4f",.)) %>% dplyr::select(-statistic) %>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/geew_alt_240502.csv"), dec=",", sep="\t")
knitr::kable(dplyr::filter(.,term=="policonsumo2"), size=10, format="markdown", caption="GEE Models, alternative weighting (after correction for PHs)")
}| model | term | estimate | std.error | p.value | conf.low | conf.high |
|---|---|---|---|---|---|---|
| simplest, binomial | policonsumo2 | 1.10 | 0.05 | 0.0457 | 1.00 | 1.21 |
| full, binomial | policonsumo2 | 1.13 | 0.05 | 0.0135 | 1.02 | 1.24 |
| simplest, poisson | policonsumo2 | 1.03 | 0.01 | 0.0184 | 1.01 | 1.06 |
| full, poisson | policonsumo2 | 1.04 | 0.01 | 0.0064 | 1.01 | 1.07 |
Stratifying
Code
# Define possible sets of breakpoints
set.seed(2125)
samp1<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(2121)
samp2<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(2121e4)
samp3<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(2121e2)
samp4<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(2121e3)
samp5<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(212e3)
samp6<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(212e4)
samp7<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(21e4)
samp8<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(21e6)
samp9<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(21e8)
samp10<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(1e2)
samp11<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(1e3)
samp12<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(1e4)
samp13<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(1e7)
samp14<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(1e5)
samp15<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(1)
samp16<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(2)
samp17<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(3)
samp17<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(4)
samp18<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(5)
samp19<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(6)
samp20<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(217e2)
samp21<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(215e2)
samp22<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(214e2)
samp23<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(213e2)
samp24<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(212e2)
samp25<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(211e2)
samp26<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(2111e3)
samp27<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(2112e3)
samp28<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(2113e3)
samp29<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
set.seed(2114e3)
samp30<- sort(c(0,sample.int(max(data_mine_miss_restr_proc2$time)-1, 5, replace = F),max(data_mine_miss_restr_proc2$time)))
breaks_list <- list(
c(0, 10, 20, 40, 60, 80, max(data_mine_miss_restr_proc2$time)),
c(0, 15, 30, 45, 60, 75, max(data_mine_miss_restr_proc2$time)),
c(0, 15, 30, 45, 60, 80, max(data_mine_miss_restr_proc2$time)),
c(0, 20, 40, 60, 80, 100, max(data_mine_miss_restr_proc2$time)),
# Unevenly spaced, emphasizing early times
c(0, 10, 20, 30, 50, 70, max(data_mine_miss_restr_proc2$time)),
# Long tail distribution
c(0, 20, 40, 60, 80, 100, 150, max(data_mine_miss_restr_proc2$time)),
# Quantile-based intervals (ensure you have no NA in `time`)
quantile(data_mine_miss_restr_proc2$time, probs = seq(0, 1, by = 0.2), na.rm = TRUE),
samp1,
samp2,
samp3,
samp4,
samp5,
samp6,
samp7,
samp8,
samp9,
samp10,
samp11,
samp12,
samp13,
samp14,
samp15,
samp16,
samp17,
samp18,
samp19,
samp20,
samp21,
samp22,
samp23,
samp24,
samp25,
samp26,
samp27,
samp28,
samp29,
samp30
)
# Prepare to store results
results_strat <- list()
# Loop over the list of breakpoints
for (i in seq_along(breaks_list)) {
# Create the factor variable with the current set of time intervals
data_mine_miss_restr_proc2$time_interval2 <- cut(data_mine_miss_restr_proc2$time, breaks_list[[i]], include.lowest = TRUE)
# Fit the Cox model
model<-
cph(as.formula(paste0("Surv(lag_time,time,event)~
cluster(id)+
lag_tr_outcome+
lag_comp_bpsc_y3_severe+
lag_less_90d_tr1+
log_lag_dias_treat_imp_sin_na +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar +
strat(tipo_de_plan_2_mod)+
strat(time_interval2)")),
data= data_mine_miss_restr_proc2 %>% data.table::as.data.table() %>% data.frame(), x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
# Perform cox.zph test
cox_test <- format(cox.zph(model)$table[nrow(cox.zph(model)$table), 1])
# Create a contingency table
contingency_table <- table(data_mine_miss_restr_proc2$time_interval2, data_mine_miss_restr_proc2$tipo_de_plan_2)
# Store results
results_strat[[i]] <- list(
AIC = AIC(model),
cox_zph = cox_test,
table = contingency_table,
fail= model$fail,
lrtest= paste0(model$stats["Model L.R."],", ",model$stats["d.f."]," p=", model$stats["P"]),
r2= model$stats["R2"]
)
}
# You can now inspect 'results' for the output from each set of breakpoints
print(results_strat)[[1]]
[[1]]$AIC
[1] 201623
[[1]]$cox_zph
[1] "132"
[[1]]$table
basic ambulatory intensive ambulatory residential
[0,10] 3697 4941 3018
(10,20] 1701 2693 1618
(20,40] 1766 3099 1774
(40,60] 911 1686 938
(60,80] 450 938 517
(80,135] 270 642 329
[[1]]$fail
[1] FALSE
[[1]]$lrtest
[1] "244.41760910835, 20 p=0"
[[1]]$r2
R2
0.0137
[[2]]
[[2]]$AIC
[1] 199180
[[2]]$cox_zph
[1] "123"
[[2]]$table
basic ambulatory intensive ambulatory residential
[0,15] 4692 6504 3976
(15,30] 1748 2892 1728
(30,45] 1018 1890 992
(45,60] 617 1133 652
(60,75] 364 762 411
(75,135] 356 818 435
[[2]]$fail
[1] FALSE
[[2]]$lrtest
[1] "236.569110802346, 20 p=0"
[[2]]$r2
R2
0.0133
[[3]]
[[3]]$AIC
[1] 199350
[[3]]$cox_zph
[1] "136"
[[3]]$table
basic ambulatory intensive ambulatory residential
[0,15] 4692 6504 3976
(15,30] 1748 2892 1728
(30,45] 1018 1890 992
(45,60] 617 1133 652
(60,80] 450 938 517
(80,135] 270 642 329
[[3]]$fail
[1] FALSE
[[3]]$lrtest
[1] "234.629002713889, 20 p=0"
[[3]]$r2
R2
0.0132
[[4]]
[[4]]$AIC
[1] 205132
[[4]]$cox_zph
[1] "149"
[[4]]$table
basic ambulatory intensive ambulatory residential
[0,20] 5398 7634 4636
(20,40] 1766 3099 1774
(40,60] 911 1686 938
(60,80] 450 938 517
(80,100] 189 472 246
(100,135] 81 170 83
[[4]]$fail
[1] FALSE
[[4]]$lrtest
[1] "266.667155946809, 20 p=0"
[[4]]$r2
R2
0.015
[[5]]
[[5]]$AIC
[1] 198563
[[5]]$cox_zph
[1] "177"
[[5]]$table
basic ambulatory intensive ambulatory residential
[0,10] 3697 4941 3018
(10,20] 1701 2693 1618
(20,30] 1042 1762 1068
(30,50] 1246 2306 1231
(50,70] 660 1275 710
(70,135] 449 1022 549
[[5]]$fail
[1] FALSE
[[5]]$lrtest
[1] "263.532100785029, 20 p=0"
[[5]]$r2
R2
0.0148
[[6]]
[[6]]$AIC
[1] 205132
[[6]]$cox_zph
[1] "149"
[[6]]$table
basic ambulatory intensive ambulatory residential
[0,20] 5398 7634 4636
(20,40] 1766 3099 1774
(40,60] 911 1686 938
(60,80] 450 938 517
(80,100] 189 472 246
(100,135] 81 170 83
(135,150] 0 0 0
[[6]]$fail
[1] FALSE
[[6]]$lrtest
[1] "266.667155946809, 20 p=0"
[[6]]$r2
R2
0.015
[[7]]
[[7]]$AIC
[1] 210729
[[7]]$cox_zph
[1] "646"
[[7]]$table
basic ambulatory intensive ambulatory residential
[0,4.85] 1887 2634 1684
(4.85,11] 2030 2648 1529
(11,22.1] 1747 2777 1658
(22.1,42.1] 1631 2899 1668
(42.1,135] 1500 3041 1655
[[7]]$fail
[1] FALSE
[[7]]$lrtest
[1] "678.659228396689, 20 p=0"
[[7]]$r2
R2
0.0377
[[8]]
[[8]]$AIC
[1] 236876
[[8]]$cox_zph
[1] "273"
[[8]]$table
basic ambulatory intensive ambulatory residential
[0,74] 8424 13137 7734
(74,90] 222 496 284
(90,108] 117 288 130
(108,117] 26 63 37
(117,128] 6 13 9
(128,135] 0 2 0
[[8]]$fail
[1] FALSE
[[8]]$lrtest
[1] "2002.72773700504, 20 p=0"
[[8]]$r2
R2
0.107
[[9]]
[[9]]$AIC
[1] 220361
[[9]]$cox_zph
[1] "289"
[[9]]$table
basic ambulatory intensive ambulatory residential
[0,18] 5132 7218 4392
(18,68] 3174 5664 3202
(68,73] 93 215 119
(73,80] 126 260 152
(80,119] 266 632 325
(119,135] 4 10 4
[[9]]$fail
[1] FALSE
[[9]]$lrtest
[1] "821.194660197216, 20 p=0"
[[9]]$r2
R2
0.0454
[[10]]
[[10]]$AIC
[1] 202224
[[10]]$cox_zph
[1] "158"
[[10]]$table
basic ambulatory intensive ambulatory residential
[0,15] 4692 6504 3976
(15,25] 1270 2044 1211
(25,50] 1724 3154 1748
(50,70] 660 1275 710
(70,85] 248 523 304
(85,135] 201 499 245
[[10]]$fail
[1] FALSE
[[10]]$lrtest
[1] "242.145208129747, 20 p=0"
[[10]]$r2
R2
0.0136
[[11]]
[[11]]$AIC
[1] 213893
[[11]]$cox_zph
[1] "211"
[[11]]$table
basic ambulatory intensive ambulatory residential
[0,1] 88 177 402
(1,24] 5779 8203 4691
(24,54] 1986 3617 2021
(54,70] 493 980 531
(70,125] 447 1019 549
(125,135] 2 3 0
[[11]]$fail
[1] FALSE
[[11]]$lrtest
[1] "450.440814013651, 20 p=0"
[[11]]$r2
R2
0.0252
[[12]]
[[12]]$AIC
[1] 220078
[[12]]$cox_zph
[1] "225"
[[12]]$table
basic ambulatory intensive ambulatory residential
[0,1] 88 177 402
(1,39] 7005 10427 5944
(39,64] 1100 2070 1126
(64,90] 453 959 546
(90,122] 147 360 176
(122,135] 2 6 0
[[12]]$fail
[1] FALSE
[[12]]$lrtest
[1] "567.485317411018, 20 p=0"
[[12]]$r2
R2
0.0316
[[13]]
[[13]]$AIC
[1] 224070
[[13]]$cox_zph
[1] "223"
[[13]]$table
basic ambulatory intensive ambulatory residential
[0,42] 7292 10953 6531
(42,78] 1198 2340 1295
(78,84] 90 182 104
(84,126] 215 521 264
(126,128] 0 1 0
(128,135] 0 2 0
[[13]]$fail
[1] FALSE
[[13]]$lrtest
[1] "892.089426955208, 20 p=0"
[[13]]$r2
R2
0.0492
[[14]]
[[14]]$AIC
[1] 231083
[[14]]$cox_zph
[1] "244"
[[14]]$table
basic ambulatory intensive ambulatory residential
[0,11] 3917 5282 3213
(11,78] 4573 8011 4613
(78,86] 115 230 143
(86,119] 186 466 221
(119,129] 4 9 4
(129,135] 0 1 0
[[14]]$fail
[1] FALSE
[[14]]$lrtest
[1] "1732.33594561392, 20 p=0"
[[14]]$r2
R2
0.0934
[[15]]
[[15]]$AIC
[1] 222381
[[15]]$cox_zph
[1] "241"
[[15]]$table
basic ambulatory intensive ambulatory residential
[0,15] 4692 6504 3976
(15,65] 3536 6218 3521
(65,90] 418 911 521
(90,107] 112 275 126
(107,120] 34 81 50
(120,135] 3 10 0
[[15]]$fail
[1] FALSE
[[15]]$lrtest
[1] "957.269989653141, 20 p=0"
[[15]]$r2
R2
0.0527
[[16]]
[[16]]$AIC
[1] 216166
[[16]]$cox_zph
[1] "317"
[[16]]$table
basic ambulatory intensive ambulatory residential
[0,29] 6350 9232 5603
(29,51] 1385 2542 1372
(51,99] 974 2042 1129
(99,110] 61 125 56
(110,131] 25 57 34
(131,135] 0 1 0
[[16]]$fail
[1] FALSE
[[16]]$lrtest
[1] "526.069504202693, 20 p=0"
[[16]]$r2
R2
0.0293
[[17]]
[[17]]$AIC
[1] 241240
[[17]]$cox_zph
[1] "336"
[[17]]$table
basic ambulatory intensive ambulatory residential
[0,88] 8623 13584 7988
(88,93] 47 113 58
(93,110] 100 244 114
(110,129] 25 57 34
(129,133] 0 0 0
(133,135] 0 1 0
[[17]]$fail
[1] FALSE
[[17]]$lrtest
[1] "2685.26883503475, 20 p=0"
[[17]]$r2
R2
0.141
[[18]]
[[18]]$AIC
[1] 227090
[[18]]$cox_zph
[1] "240"
[[18]]$table
basic ambulatory intensive ambulatory residential
[0,4] 1474 2139 1422
(4,55] 6419 9938 5742
(55,70] 453 900 481
(70,102] 387 869 477
(102,112] 47 111 43
(112,135] 15 42 29
[[18]]$fail
[1] FALSE
[[18]]$lrtest
[1] "1063.35690842738, 20 p=0"
[[18]]$r2
R2
0.0584
[[19]]
[[19]]$AIC
[1] 204946
[[19]]$cox_zph
[1] "222"
[[19]]$table
basic ambulatory intensive ambulatory residential
[0,29] 6350 9232 5603
(29,43] 998 1837 989
(43,51] 387 705 383
(51,68] 571 1108 619
(68,88] 317 702 394
(88,135] 172 415 206
[[19]]$fail
[1] FALSE
[[19]]$lrtest
[1] "309.440397133556, 20 p=0"
[[19]]$r2
R2
0.0174
[[20]]
[[20]]$AIC
[1] 207600
[[20]]$cox_zph
[1] "163"
[[20]]$table
basic ambulatory intensive ambulatory residential
[0,9] 3455 4593 2807
(9,30] 2985 4803 2897
(30,45] 1018 1890 992
(45,79] 1052 2043 1148
(79,84] 70 146 86
(84,135] 215 524 264
[[20]]$fail
[1] FALSE
[[20]]$lrtest
[1] "418.016287462291, 20 p=0"
[[20]]$r2
R2
0.0234
[[21]]
[[21]]$AIC
[1] 227986
[[21]]$cox_zph
[1] "172"
[[21]]$table
basic ambulatory intensive ambulatory residential
[0,6] 2395 3270 2033
(6,9] 1060 1323 774
(9,70] 4891 8384 4838
(70,80] 179 380 220
(80,92] 134 317 175
(92,135] 136 325 154
[[21]]$fail
[1] FALSE
[[21]]$lrtest
[1] "1537.47392012289, 20 p=0"
[[21]]$r2
R2
0.0833
[[22]]
[[22]]$AIC
[1] 228537
[[22]]$cox_zph
[1] "232"
[[22]]$table
basic ambulatory intensive ambulatory residential
[0,54] 7853 11997 7114
(54,76] 609 1226 666
(76,106] 291 678 360
(106,107] 5 7 4
(107,131] 37 90 50
(131,135] 0 1 0
[[22]]$fail
[1] FALSE
[[22]]$lrtest
[1] "1075.1562108639, 20 p=0"
[[22]]$r2
R2
0.059
[[23]]
[[23]]$AIC
[1] 209837
[[23]]$cox_zph
[1] "168"
[[23]]$table
basic ambulatory intensive ambulatory residential
[0,14] 4509 6237 3819
(14,43] 2839 4832 2773
(43,51] 387 705 383
(51,68] 571 1108 619
(68,129] 489 1116 600
(129,135] 0 1 0
[[23]]$fail
[1] FALSE
[[23]]$lrtest
[1] "453.468024794711, 20 p=0"
[[23]]$r2
R2
0.0253
[[24]]
[[24]]$AIC
[1] 209169
[[24]]$cox_zph
[1] "224"
[[24]]$table
basic ambulatory intensive ambulatory residential
[0,5] 1947 2708 1720
(5,20] 3451 4926 2916
(20,36] 1499 2599 1538
(36,74] 1527 2904 1560
(74,107] 334 771 410
(107,135] 37 91 50
[[24]]$fail
[1] FALSE
[[24]]$lrtest
[1] "440.86968850813, 20 p=0"
[[24]]$r2
R2
0.0246
[[25]]
[[25]]$AIC
[1] 225610
[[25]]$cox_zph
[1] "260"
[[25]]$table
basic ambulatory intensive ambulatory residential
[0,3] 915 1436 1091
(3,51] 6820 10338 5884
(51,71] 629 1242 700
(71,75] 75 165 84
(75,115] 347 789 421
(115,135] 9 29 14
[[25]]$fail
[1] FALSE
[[25]]$lrtest
[1] "908.308044379839, 20 p=0"
[[25]]$r2
R2
0.0501
[[26]]
[[26]]$AIC
[1] 220425
[[26]]$cox_zph
[1] "244"
[[26]]$table
basic ambulatory intensive ambulatory residential
[0,41] 7224 10837 6468
(41,66] 1024 1948 1067
(66,71] 116 231 140
(71,107] 394 892 469
(107,121] 35 82 50
(121,135] 2 9 0
[[26]]$fail
[1] FALSE
[[26]]$lrtest
[1] "695.022012121801, 20 p=0"
[[26]]$r2
R2
0.0386
[[27]]
[[27]]$AIC
[1] 218142
[[27]]$cox_zph
[1] "152"
[[27]]$table
basic ambulatory intensive ambulatory residential
[0,10] 3697 4941 3018
(10,53] 4118 6978 4055
(53,60] 260 500 275
(60,78] 415 874 478
(78,100] 224 536 285
(100,135] 81 170 83
[[27]]$fail
[1] FALSE
[[27]]$lrtest
[1] "743.24174347092, 20 p=0"
[[27]]$r2
R2
0.0412
[[28]]
[[28]]$AIC
[1] 218161
[[28]]$cox_zph
[1] "252"
[[28]]$table
basic ambulatory intensive ambulatory residential
[0,33] 6697 9826 5947
(33,36] 200 407 227
(36,76] 1565 2990 1606
(76,77] 15 23 20
(77,122] 316 747 394
(122,135] 2 6 0
[[28]]$fail
[1] FALSE
[[28]]$lrtest
[1] "658.233573928737, 20 p=0"
[[28]]$r2
R2
0.0366
[[29]]
[[29]]$AIC
[1] 216388
[[29]]$cox_zph
[1] "236"
[[29]]$table
basic ambulatory intensive ambulatory residential
[0,36] 6897 10233 6174
(36,52] 873 1621 853
(52,86] 835 1669 942
(86,102] 128 323 153
(102,132] 62 152 72
(132,135] 0 1 0
[[29]]$fail
[1] FALSE
[[29]]$lrtest
[1] "513.853596449771, 20 p=0"
[[29]]$r2
R2
0.0287
[[30]]
[[30]]$AIC
[1] 197479
[[30]]$cox_zph
[1] "396"
[[30]]$table
basic ambulatory intensive ambulatory residential
[0,13] 4335 5941 3619
(13,17] 657 1039 637
(17,28] 1271 2094 1259
(28,36] 634 1159 659
(36,52] 873 1621 853
(52,135] 1025 2145 1167
[[30]]$fail
[1] FALSE
[[30]]$lrtest
[1] "316.936203734804, 20 p=0"
[[30]]$r2
R2
0.0178
[[31]]
[[31]]$AIC
[1] 224694
[[31]]$cox_zph
[1] "253"
[[31]]$table
basic ambulatory intensive ambulatory residential
[0,50] 7686 11702 6935
(50,62] 449 838 469
(62,96] 555 1202 672
(96,106] 63 159 64
(106,116] 34 77 44
(116,135] 8 21 10
[[31]]$fail
[1] FALSE
[[31]]$lrtest
[1] "908.547777739586, 20 p=0"
[[31]]$r2
R2
0.0501
[[32]]
[[32]]$AIC
[1] 232440
[[32]]$cox_zph
[1] "249"
[[32]]$table
basic ambulatory intensive ambulatory residential
[0,62] 8135 12540 7404
(62,73] 264 557 309
(73,113] 382 861 455
(113,120] 11 31 26
(120,129] 3 9 0
(129,135] 0 1 0
[[32]]$fail
[1] FALSE
[[32]]$lrtest
[1] "1438.46574364958, 20 p=0"
[[32]]$r2
R2
0.0782
[[33]]
[[33]]$AIC
[1] 229114
[[33]]$cox_zph
[1] "262"
[[33]]$table
basic ambulatory intensive ambulatory residential
[0,55] 7893 12077 7164
(55,81] 645 1318 716
(81,82] 13 35 19
(82,107] 207 478 245
(107,124] 35 88 50
(124,135] 2 3 0
[[33]]$fail
[1] FALSE
[[33]]$lrtest
[1] "1131.15766450859, 20 p=0"
[[33]]$r2
R2
0.062
[[34]]
[[34]]$AIC
[1] 207443
[[34]]$cox_zph
[1] "324"
[[34]]$table
basic ambulatory intensive ambulatory residential
[0,16] 4837 6759 4118
(16,38] 2194 3717 2174
(38,43] 317 593 300
(43,51] 387 705 383
(51,132] 1060 2224 1219
(132,135] 0 1 0
[[34]]$fail
[1] FALSE
[[34]]$lrtest
[1] "488.713657127577, 20 p=0"
[[34]]$r2
R2
0.0273
[[35]]
[[35]]$AIC
[1] 215882
[[35]]$cox_zph
[1] "685"
[[35]]$table
basic ambulatory intensive ambulatory residential
[0,8] 3151 4168 2554
(8,28] 3112 4906 2961
(28,35] 568 1010 582
(35,112] 1949 3873 2068
(112,114] 5 9 10
(114,135] 10 33 19
[[35]]$fail
[1] FALSE
[[35]]$lrtest
[1] "872.879308136587, 20 p=0"
[[35]]$r2
R2
0.0482
[[36]]
[[36]]$AIC
[1] 226868
[[36]]$cox_zph
[1] "466"
[[36]]$table
basic ambulatory intensive ambulatory residential
[0,1] 88 177 402
(1,34] 6681 9783 5633
(34,98] 1932 3827 2062
(98,128] 94 210 97
(128,132] 0 1 0
(132,135] 0 1 0
[[36]]$fail
[1] FALSE
[[36]]$lrtest
[1] "1064.25334982859, 20 p=0"
[[36]]$r2
R2
0.0584
[[37]]
[[37]]$AIC
[1] 219085
[[37]]$cox_zph
[1] "185"
[[37]]$table
basic ambulatory intensive ambulatory residential
[0,4] 1474 2139 1422
(4,10] 2223 2802 1596
(10,45] 3761 6345 3678
(45,75] 981 1895 1063
(75,124] 354 815 435
(124,135] 2 3 0
[[37]]$fail
[1] FALSE
[[37]]$lrtest
[1] "697.807880941953, 20 p=0"
[[37]]$r2
R2
0.0387 Code
print(results_strat[[2]])$AIC
[1] 199180
$cox_zph
[1] "123"
$table
basic ambulatory intensive ambulatory residential
[0,15] 4692 6504 3976
(15,30] 1748 2892 1728
(30,45] 1018 1890 992
(45,60] 617 1133 652
(60,75] 364 762 411
(75,135] 356 818 435
$fail
[1] FALSE
$lrtest
[1] "236.569110802346, 20 p=0"
$r2
R2
0.0133 Code
#117
invisible("El original X2= 430.6824, df=18, p< 2e-16; ")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("SELECTY INTERVAL TIME TO STRATIFY")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/selected_models_strat.R")
# Create a factor variable with time intervals
data_mine_miss_restr_proc2$time_interval3 <- cut(data_mine_miss_restr_proc2$time,
breaks_list[[17]][c(1:3,7)],
include.lowest = TRUE)
#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:
model_after_time_strat<-
cph(Surv(lag_time,time,event)~
cluster(id)+
lag_tr_outcome+
lag_comp_bpsc_y3_severe+
lag_less_90d_tr1+
log_lag_dias_treat_imp_sin_na +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar +
strat(tipo_de_plan_2_mod)+
strat(time_interval3),
data= data_mine_miss_restr_proc2 %>% data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
#
model_after_time_stratFrequencies of Missing Values Due to Each Variable
Surv(lag_time, time, event) cluster(id)
13317 0
lag_tr_outcome lag_comp_bpsc_y3_severe
0 0
lag_less_90d_tr1 log_lag_dias_treat_imp_sin_na
0 13317
lag_policonsumo2 edad_al_ing_1
0 0
ano_nac_corr susinidum_oh
0 0
susinidum_coc susinidum_pbc
0 0
susinidum_mar psycom_dum_study
0 0
psycom_dum_with freq_cons_dum_5day
0 0
cond_oc_dum_2inact cond_oc_dum_3unemp
0 0
susprindum_oh susprindum_coc
0 0
susprindum_pbc susprindum_mar
0 0
tipo_de_plan_2_mod time_interval3
0 0
Cox Proportional Hazards Model
cph(formula = Surv(lag_time, time, event) ~ cluster(id) + lag_tr_outcome +
lag_comp_bpsc_y3_severe + lag_less_90d_tr1 + log_lag_dias_treat_imp_sin_na +
lag_policonsumo2 + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_study +
psycom_dum_with + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar + strat(tipo_de_plan_2_mod) + strat(time_interval3),
data = data_mine_miss_restr_proc2 %>% data.table::as.data.table() %>%
data.frame(), x = TRUE, y = TRUE, iter.max = 250 * 4,
tol = 1e-06, surv = TRUE)
Status
Stratum No Event
tipo_de_plan_2_mod=basic ambulatory.time_interval3=[0,88] 0
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3=[0,88] 0
tipo_de_plan_2_mod=GP residential.time_interval3=[0,88] 0
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3=[0,88] 0
tipo_de_plan_2_mod=WO residential.time_interval3=[0,88] 0
tipo_de_plan_2_mod=basic ambulatory.time_interval3=(88,93] 0
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3=(88,93] 0
tipo_de_plan_2_mod=GP residential.time_interval3=(88,93] 0
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3=(88,93] 0
tipo_de_plan_2_mod=WO residential.time_interval3=(88,93] 0
tipo_de_plan_2_mod=basic ambulatory.time_interval3=(93,135] 0
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3=(93,135] 0
tipo_de_plan_2_mod=GP residential.time_interval3=(93,135] 0
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3=(93,135] 0
tipo_de_plan_2_mod=WO residential.time_interval3=(93,135] 0
Status
Stratum Event
tipo_de_plan_2_mod=basic ambulatory.time_interval3=[0,88] 5356
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3=[0,88] 6212
tipo_de_plan_2_mod=GP residential.time_interval3=[0,88] 2895
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3=[0,88] 964
tipo_de_plan_2_mod=WO residential.time_interval3=[0,88] 1451
tipo_de_plan_2_mod=basic ambulatory.time_interval3=(88,93] 77
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3=(88,93] 74
tipo_de_plan_2_mod=GP residential.time_interval3=(88,93] 44
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3=(88,93] 10
tipo_de_plan_2_mod=WO residential.time_interval3=(88,93] 13
tipo_de_plan_2_mod=basic ambulatory.time_interval3=(93,135] 200
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3=(93,135] 213
tipo_de_plan_2_mod=GP residential.time_interval3=(93,135] 87
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3=(93,135] 41
tipo_de_plan_2_mod=WO residential.time_interval3=(93,135] 34
Model Tests Discrimination
Indexes
Obs 17671 LR chi2 2722.85 R2 0.143
Events 17671 d.f. 20 R2(20,17671)0.142
Center 402 Pr(> chi2) 0.0000 Dxy 0.256
Score chi2 2731.11
Pr(> chi2) 0.0000
Coef S.E. Wald Z Pr(>|Z|)
lag_tr_outcome 0.1374 0.0182 7.55 <0.0001
lag_comp_bpsc_y3_severe 0.0508 0.0169 3.00 0.0027
lag_less_90d_tr1 0.1048 0.0252 4.16 <0.0001
log_lag_dias_treat_imp_sin_na -0.0197 0.0110 -1.79 0.0733
lag_policonsumo2 -0.0206 0.0191 -1.07 0.2828
edad_al_ing_1 0.1987 0.0038 52.28 <0.0001
ano_nac_corr 0.1997 0.0037 53.44 <0.0001
susinidum_oh 0.0440 0.0483 0.91 0.3623
susinidum_coc 0.1127 0.0595 1.89 0.0582
susinidum_pbc 0.1409 0.0545 2.58 0.0098
susinidum_mar 0.1610 0.0492 3.27 0.0011
psycom_dum_study 0.0132 0.0231 0.57 0.5691
psycom_dum_with 0.0145 0.0164 0.89 0.3759
freq_cons_dum_5day 0.0094 0.0160 0.59 0.5556
cond_oc_dum_2inact 0.0757 0.0218 3.47 0.0005
cond_oc_dum_3unemp 0.0841 0.0176 4.79 <0.0001
susprindum_oh -0.0798 0.0679 -1.18 0.2398
susprindum_coc -0.0757 0.0687 -1.10 0.2708
susprindum_pbc -0.0578 0.0675 -0.86 0.3917
susprindum_mar -0.0801 0.0744 -1.08 0.2814
Code
# Cox Proportional Hazards Model
#
# cph(formula = Surv(lag_time, time, event) ~ cluster(id) + lag_tr_outcome +
# lag_comp_bpsc_y3_severe + lag_less_90d_tr1 + log_lag_dias_treat_imp_sin_na +
# lag_policonsumo2 + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
# susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_study +
# psycom_dum_with + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar + strat(tipo_de_plan_2_mod) + strat(time_interval3),
# data = data_mine_miss_restr_proc2 %>% data.table::as.data.table() %>%
# data.frame(), x = TRUE, y = TRUE, iter.max = 250 * 4,
# tol = 1e-06, surv = TRUE)
#
#
# Status
# Stratum No Event Event
# tipo_de_plan_2_mod=basic ambulatory.time_interval3=[0,88] 0 5356
# tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3=[0,88] 0 6212
# tipo_de_plan_2_mod=GP residential.time_interval3=[0,88] 0 2895
# tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3=[0,88] 0 964
# tipo_de_plan_2_mod=WO residential.time_interval3=[0,88] 0 1451
# tipo_de_plan_2_mod=basic ambulatory.time_interval3=(88,93] 0 77
# tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3=(88,93] 0 74
# tipo_de_plan_2_mod=GP residential.time_interval3=(88,93] 0 44
# tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3=(88,93] 0 10
# tipo_de_plan_2_mod=WO residential.time_interval3=(88,93] 0 13
# tipo_de_plan_2_mod=basic ambulatory.time_interval3=(93,135] 0 200
# tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3=(93,135] 0 213
# tipo_de_plan_2_mod=GP residential.time_interval3=(93,135] 0 87
# tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3=(93,135] 0 41
# tipo_de_plan_2_mod=WO residential.time_interval3=(93,135] 0 34
#
# Model Tests Discrimination
# Indexes
# Obs 17671 LR chi2 2722.85 R2 0.143
# Events 17671 d.f. 20 R2(20,17671)0.142
# Center 402 Pr(> chi2) 0.0000 Dxy 0.256
# Score chi2 2731.11
# Pr(> chi2) 0.0000
#
# Coef S.E. Wald Z Pr(>|Z|)
# lag_tr_outcome 0.1374 0.0182 7.55 <0.0001
# lag_comp_bpsc_y3_severe 0.0508 0.0169 3.00 0.0027
# lag_less_90d_tr1 0.1048 0.0252 4.16 <0.0001
# log_lag_dias_treat_imp_sin_na -0.0197 0.0110 -1.79 0.0733
# lag_policonsumo2 -0.0206 0.0191 -1.07 0.2828
# edad_al_ing_1 0.1987 0.0038 52.28 <0.0001
# ano_nac_corr 0.1997 0.0037 53.44 <0.0001
# susinidum_oh 0.0440 0.0483 0.91 0.3623
# susinidum_coc 0.1127 0.0595 1.89 0.0582
# susinidum_pbc 0.1409 0.0545 2.58 0.0098
# susinidum_mar 0.1610 0.0492 3.27 0.0011
# psycom_dum_study 0.0132 0.0231 0.57 0.5691
# psycom_dum_with 0.0145 0.0164 0.89 0.3759
# freq_cons_dum_5day 0.0094 0.0160 0.59 0.5556
# cond_oc_dum_2inact 0.0757 0.0218 3.47 0.0005
# cond_oc_dum_3unemp 0.0841 0.0176 4.79 <0.0001
# susprindum_oh -0.0798 0.0679 -1.18 0.2398
# susprindum_coc -0.0757 0.0687 -1.10 0.2708
# susprindum_pbc -0.0578 0.0675 -0.86 0.3917
# susprindum_mar -0.0801 0.0744 -1.08 0.2814
cox.zph(model_after_time_strat) chisq df p
lag_tr_outcome 60.354 1 7.9e-15
lag_comp_bpsc_y3_severe 20.528 1 5.9e-06
lag_less_90d_tr1 151.524 1 < 2e-16
log_lag_dias_treat_imp_sin_na 160.971 1 < 2e-16
lag_policonsumo2 5.712 1 0.01685
edad_al_ing_1 31.506 1 2.0e-08
ano_nac_corr 19.244 1 1.2e-05
susinidum_oh 35.309 1 2.8e-09
susinidum_coc 0.931 1 0.33456
susinidum_pbc 10.487 1 0.00120
susinidum_mar 17.059 1 3.6e-05
psycom_dum_study 51.800 1 6.1e-13
psycom_dum_with 18.230 1 2.0e-05
freq_cons_dum_5day 3.577 1 0.05860
cond_oc_dum_2inact 3.270 1 0.07055
cond_oc_dum_3unemp 7.459 1 0.00631
susprindum_oh 16.966 1 3.8e-05
susprindum_coc 0.186 1 0.66631
susprindum_pbc 12.610 1 0.00038
susprindum_mar 1.337 1 0.24758
GLOBAL 323.981 20 < 2e-16Code
# chisq df p
# lag_tr_outcome 60.354 1 7.9e-15
# lag_comp_bpsc_y3_severe 20.528 1 5.9e-06
# lag_less_90d_tr1 151.524 1 < 2e-16
# log_lag_dias_treat_imp_sin_na 160.971 1 < 2e-16
# lag_policonsumo2 5.712 1 0.01685
# edad_al_ing_1 31.506 1 2.0e-08
# ano_nac_corr 19.244 1 1.2e-05
# susinidum_oh 35.309 1 2.8e-09
# susinidum_coc 0.931 1 0.33456
# susinidum_pbc 10.487 1 0.00120
# susinidum_mar 17.059 1 3.6e-05
# psycom_dum_study 51.800 1 6.1e-13
# psycom_dum_with 18.230 1 2.0e-05
# freq_cons_dum_5day 3.577 1 0.05860
# cond_oc_dum_2inact 3.270 1 0.07055
# cond_oc_dum_3unemp 7.459 1 0.00631
# susprindum_oh 16.966 1 3.8e-05
# susprindum_coc 0.186 1 0.66631
# susprindum_pbc 12.610 1 0.00038
# susprindum_mar 1.337 1 0.24758
# GLOBAL 323.981 20 < 2e-16
table(data_mine_miss_restr_proc2$time_interval3, data_mine_miss_restr_proc2$tipo_de_plan_2_mod)
basic ambulatory GP intensive ambulatory GP residential
[0,88] 9716 11210 5073
(88,93] 77 74 44
(93,135] 200 213 87
WO intensive ambulatory WO residential
[0,88] 1709 2487
(88,93] 10 13
(93,135] 41 34Code
# basic ambulatory GP intensive ambulatory GP residential WO intensive ambulatory WO residential
# [0,88] 9716 11210 5073 1709 2487
# (88,93] 77 74 44 10 13
# (93,135] 200 213 87 41 34
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
data_mine_miss_restr_proc2$time_interval3_alt <- cut(data_mine_miss_restr_proc2$time,
breaks_list[[5]],
include.lowest = TRUE)
model_after_time_strat_alt<-
cph(Surv(lag_time,time,event)~
cluster(id)+
lag_tr_outcome+
lag_comp_bpsc_y3_severe+
lag_less_90d_tr1+
log_lag_dias_treat_imp_sin_na +
lag_policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar +
strat(tipo_de_plan_2_mod)+
strat(time_interval3_alt),
data= data_mine_miss_restr_proc2 %>% data.table::as.data.table() %>% data.frame(),
x=TRUE, y=TRUE, surv=TRUE, iter.max = 250*4, tol = 1e-6)
#
model_after_time_strat_altFrequencies of Missing Values Due to Each Variable
Surv(lag_time, time, event) cluster(id)
13317 0
lag_tr_outcome lag_comp_bpsc_y3_severe
0 0
lag_less_90d_tr1 log_lag_dias_treat_imp_sin_na
0 13317
lag_policonsumo2 edad_al_ing_1
0 0
ano_nac_corr susinidum_oh
0 0
susinidum_coc susinidum_pbc
0 0
susinidum_mar psycom_dum_study
0 0
psycom_dum_with freq_cons_dum_5day
0 0
cond_oc_dum_2inact cond_oc_dum_3unemp
0 0
susprindum_oh susprindum_coc
0 0
susprindum_pbc susprindum_mar
0 0
tipo_de_plan_2_mod time_interval3_alt
0 0
Cox Proportional Hazards Model
cph(formula = Surv(lag_time, time, event) ~ cluster(id) + lag_tr_outcome +
lag_comp_bpsc_y3_severe + lag_less_90d_tr1 + log_lag_dias_treat_imp_sin_na +
lag_policonsumo2 + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_study +
psycom_dum_with + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar + strat(tipo_de_plan_2_mod) + strat(time_interval3_alt),
data = data_mine_miss_restr_proc2 %>% data.table::as.data.table() %>%
data.frame(), x = TRUE, y = TRUE, iter.max = 250 * 4,
tol = 1e-06, surv = TRUE)
Status
Stratum No Event
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=[0,10] 0
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=[0,10] 0
tipo_de_plan_2_mod=GP residential.time_interval3_alt=[0,10] 0
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=[0,10] 0
tipo_de_plan_2_mod=WO residential.time_interval3_alt=[0,10] 0
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(10,20] 0
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(10,20] 0
tipo_de_plan_2_mod=GP residential.time_interval3_alt=(10,20] 0
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(10,20] 0
tipo_de_plan_2_mod=WO residential.time_interval3_alt=(10,20] 0
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(20,30] 0
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(20,30] 0
tipo_de_plan_2_mod=GP residential.time_interval3_alt=(20,30] 0
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(20,30] 0
tipo_de_plan_2_mod=WO residential.time_interval3_alt=(20,30] 0
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(30,50] 0
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(30,50] 0
tipo_de_plan_2_mod=GP residential.time_interval3_alt=(30,50] 0
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(30,50] 0
tipo_de_plan_2_mod=WO residential.time_interval3_alt=(30,50] 0
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(50,70] 0
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(50,70] 0
tipo_de_plan_2_mod=GP residential.time_interval3_alt=(50,70] 0
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(50,70] 0
tipo_de_plan_2_mod=WO residential.time_interval3_alt=(50,70] 0
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(70,135] 0
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(70,135] 0
tipo_de_plan_2_mod=GP residential.time_interval3_alt=(70,135] 0
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(70,135] 0
tipo_de_plan_2_mod=WO residential.time_interval3_alt=(70,135] 0
Status
Stratum Event
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=[0,10] 313
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=[0,10] 427
tipo_de_plan_2_mod=GP residential.time_interval3_alt=[0,10] 306
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=[0,10] 55
tipo_de_plan_2_mod=WO residential.time_interval3_alt=[0,10] 178
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(10,20] 1040
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(10,20] 1288
tipo_de_plan_2_mod=GP residential.time_interval3_alt=(10,20] 673
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(10,20] 223
tipo_de_plan_2_mod=WO residential.time_interval3_alt=(10,20] 322
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(20,30] 1073
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(20,30] 1311
tipo_de_plan_2_mod=GP residential.time_interval3_alt=(20,30] 578
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(20,30] 216
tipo_de_plan_2_mod=WO residential.time_interval3_alt=(20,30] 310
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(30,50] 1535
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(30,50] 1774
tipo_de_plan_2_mod=GP residential.time_interval3_alt=(30,50] 752
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(30,50] 265
tipo_de_plan_2_mod=WO residential.time_interval3_alt=(30,50] 370
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(50,70] 973
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(50,70] 952
tipo_de_plan_2_mod=GP residential.time_interval3_alt=(50,70] 392
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(50,70] 147
tipo_de_plan_2_mod=WO residential.time_interval3_alt=(50,70] 178
tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(70,135] 699
tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(70,135] 747
tipo_de_plan_2_mod=GP residential.time_interval3_alt=(70,135] 325
tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(70,135] 109
tipo_de_plan_2_mod=WO residential.time_interval3_alt=(70,135] 140
Model Tests Discrimination
Indexes
Obs 17671 LR chi2 263.53 R2 0.015
Events 17671 d.f. 20 R2(20,17671)0.014
Center 105 Pr(> chi2) 0.0000 Dxy 0.251
Score chi2 263.20
Pr(> chi2) 0.0000
Coef S.E. Wald Z Pr(>|Z|)
lag_tr_outcome 0.0448 0.0190 2.36 0.0183
lag_comp_bpsc_y3_severe 0.0238 0.0170 1.40 0.1629
lag_less_90d_tr1 0.0374 0.0227 1.65 0.0993
log_lag_dias_treat_imp_sin_na -0.0289 0.0084 -3.44 0.0006
lag_policonsumo2 -0.0071 0.0201 -0.35 0.7234
edad_al_ing_1 0.0520 0.0038 13.68 <0.0001
ano_nac_corr 0.0524 0.0038 13.96 <0.0001
susinidum_oh -0.0306 0.0524 -0.58 0.5591
susinidum_coc 0.0607 0.0617 0.98 0.3256
susinidum_pbc -0.0096 0.0573 -0.17 0.8671
susinidum_mar -0.0109 0.0529 -0.21 0.8361
psycom_dum_study 0.0047 0.0220 0.21 0.8311
psycom_dum_with -0.0208 0.0169 -1.23 0.2190
freq_cons_dum_5day 0.0204 0.0161 1.26 0.2067
cond_oc_dum_2inact -0.0145 0.0222 -0.65 0.5137
cond_oc_dum_3unemp 0.0231 0.0178 1.30 0.1938
susprindum_oh 0.0811 0.0694 1.17 0.2423
susprindum_coc 0.0585 0.0700 0.84 0.4034
susprindum_pbc 0.0645 0.0684 0.94 0.3454
susprindum_mar 0.0681 0.0745 0.91 0.3607
Code
# Cox Proportional Hazards Model
#
# cph(formula = Surv(lag_time, time, event) ~ cluster(id) + lag_tr_outcome +
# lag_comp_bpsc_y3_severe + lag_less_90d_tr1 + log_lag_dias_treat_imp_sin_na +
# lag_policonsumo2 + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
# susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_study +
# psycom_dum_with + freq_cons_dum_5day + cond_oc_dum_2inact +
# cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
# susprindum_mar + strat(tipo_de_plan_2_mod) + strat(time_interval3_alt),
# data = data_mine_miss_restr_proc2 %>% data.table::as.data.table() %>%
# data.frame(), x = TRUE, y = TRUE, iter.max = 250 * 4,
# tol = 1e-06, surv = TRUE)
#
#
# Status
# Stratum No Event Event
# tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=[0,10] 0 313
# tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=[0,10] 0 427
# tipo_de_plan_2_mod=GP residential.time_interval3_alt=[0,10] 0 306
# tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=[0,10] 0 55
# tipo_de_plan_2_mod=WO residential.time_interval3_alt=[0,10] 0 178
# tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(10,20] 0 1040
# tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(10,20] 0 1288
# tipo_de_plan_2_mod=GP residential.time_interval3_alt=(10,20] 0 673
# tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(10,20] 0 223
# tipo_de_plan_2_mod=WO residential.time_interval3_alt=(10,20] 0 322
# tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(20,30] 0 1073
# tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(20,30] 0 1311
# tipo_de_plan_2_mod=GP residential.time_interval3_alt=(20,30] 0 578
# tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(20,30] 0 216
# tipo_de_plan_2_mod=WO residential.time_interval3_alt=(20,30] 0 310
# tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(30,50] 0 1535
# tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(30,50] 0 1774
# tipo_de_plan_2_mod=GP residential.time_interval3_alt=(30,50] 0 752
# tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(30,50] 0 265
# tipo_de_plan_2_mod=WO residential.time_interval3_alt=(30,50] 0 370
# tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(50,70] 0 973
# tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(50,70] 0 952
# tipo_de_plan_2_mod=GP residential.time_interval3_alt=(50,70] 0 392
# tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(50,70] 0 147
# tipo_de_plan_2_mod=WO residential.time_interval3_alt=(50,70] 0 178
# tipo_de_plan_2_mod=basic ambulatory.time_interval3_alt=(70,135] 0 699
# tipo_de_plan_2_mod=GP intensive ambulatory.time_interval3_alt=(70,135] 0 747
# tipo_de_plan_2_mod=GP residential.time_interval3_alt=(70,135] 0 325
# tipo_de_plan_2_mod=WO intensive ambulatory.time_interval3_alt=(70,135] 0 109
# tipo_de_plan_2_mod=WO residential.time_interval3_alt=(70,135] 0 140
#
# Model Tests Discrimination
# Indexes
# Obs 17671 LR chi2 263.53 R2 0.015
# Events 17671 d.f. 20 R2(20,17671)0.014
# Center 105 Pr(> chi2) 0.0000 Dxy 0.251
# Score chi2 263.20
# Pr(> chi2) 0.0000
#
# Coef S.E. Wald Z Pr(>|Z|)
# lag_tr_outcome 0.0448 0.0190 2.36 0.0183
# lag_comp_bpsc_y3_severe 0.0238 0.0170 1.40 0.1629
# lag_less_90d_tr1 0.0374 0.0227 1.65 0.0993
# log_lag_dias_treat_imp_sin_na -0.0289 0.0084 -3.44 0.0006
# lag_policonsumo2 -0.0071 0.0201 -0.35 0.7234
# edad_al_ing_1 0.0520 0.0038 13.68 <0.0001
# ano_nac_corr 0.0524 0.0038 13.96 <0.0001
# susinidum_oh -0.0306 0.0524 -0.58 0.5591
# susinidum_coc 0.0607 0.0617 0.98 0.3256
# susinidum_pbc -0.0096 0.0573 -0.17 0.8671
# susinidum_mar -0.0109 0.0529 -0.21 0.8361
# psycom_dum_study 0.0047 0.0220 0.21 0.8311
# psycom_dum_with -0.0208 0.0169 -1.23 0.2190
# freq_cons_dum_5day 0.0204 0.0161 1.26 0.2067
# cond_oc_dum_2inact -0.0145 0.0222 -0.65 0.5137
# cond_oc_dum_3unemp 0.0231 0.0178 1.30 0.1938
# susprindum_oh 0.0811 0.0694 1.17 0.2423
# susprindum_coc 0.0585 0.0700 0.84 0.4034
# susprindum_pbc 0.0645 0.0684 0.94 0.3454
# susprindum_mar 0.0681 0.0745 0.91 0.3607
cox.zph(model_after_time_strat_alt) chisq df p
lag_tr_outcome 6.8110 1 0.0091
lag_comp_bpsc_y3_severe 3.2478 1 0.0715
lag_less_90d_tr1 16.6116 1 4.6e-05
log_lag_dias_treat_imp_sin_na 32.2194 1 1.4e-08
lag_policonsumo2 0.6498 1 0.4202
edad_al_ing_1 3.4324 1 0.0639
ano_nac_corr 0.2407 1 0.6237
susinidum_oh 3.3813 1 0.0659
susinidum_coc 1.5745 1 0.2096
susinidum_pbc 1.7148 1 0.1904
susinidum_mar 0.5145 1 0.4732
psycom_dum_study 5.8375 1 0.0157
psycom_dum_with 3.0765 1 0.0794
freq_cons_dum_5day 0.2376 1 0.6259
cond_oc_dum_2inact 0.7335 1 0.3918
cond_oc_dum_3unemp 4.4843 1 0.0342
susprindum_oh 4.5222 1 0.0335
susprindum_coc 0.0816 1 0.7751
susprindum_pbc 2.2638 1 0.1324
susprindum_mar 1.6041 1 0.2053
GLOBAL 177.1396 20 < 2e-16Code
# chisq df p
# lag_tr_outcome 6.8110 1 0.0091
# lag_comp_bpsc_y3_severe 3.2478 1 0.0715
# lag_less_90d_tr1 16.6116 1 4.6e-05
# log_lag_dias_treat_imp_sin_na 32.2194 1 1.4e-08
# lag_policonsumo2 0.6498 1 0.4202
# edad_al_ing_1 3.4324 1 0.0639
# ano_nac_corr 0.2407 1 0.6237
# susinidum_oh 3.3813 1 0.0659
# susinidum_coc 1.5745 1 0.2096
# susinidum_pbc 1.7148 1 0.1904
# susinidum_mar 0.5145 1 0.4732
# psycom_dum_study 5.8375 1 0.0157
# psycom_dum_with 3.0765 1 0.0794
# freq_cons_dum_5day 0.2376 1 0.6259
# cond_oc_dum_2inact 0.7335 1 0.3918
# cond_oc_dum_3unemp 4.4843 1 0.0342
# susprindum_oh 4.5222 1 0.0335
# susprindum_coc 0.0816 1 0.7751
# susprindum_pbc 2.2638 1 0.1324
# susprindum_mar 1.6041 1 0.2053
# GLOBAL 177.1396 20 < 2e-16
table(data_mine_miss_restr_proc2$time_interval3_alt, data_mine_miss_restr_proc2$tipo_de_plan_2_mod)
basic ambulatory GP intensive ambulatory GP residential
[0,10] 3756 4289 2015
(10,20] 1787 2228 1104
(20,30] 1208 1468 612
(30,50] 1568 1813 756
(50,70] 975 952 392
(70,135] 699 747 325
WO intensive ambulatory WO residential
[0,10] 602 994
(10,20] 383 510
(20,30] 247 337
(30,50] 272 374
(50,70] 147 179
(70,135] 109 140Code
# basic ambulatory GP intensive ambulatory GP residential WO intensive ambulatory WO residential
# [0,10] 3756 4289 2015 602 994
# (10,20] 1787 2228 1104 383 510
# (20,30] 1208 1468 612 247 337
# (30,50] 1568 1813 756 272 374
# (50,70] 975 952 392 147 179
# (70,135] 699 747 325 109 140Code
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
save.image(paste0(folder_path,"20240503_avance_iiw.RData"))Apply
Code
#iiws_strat_id <- data.frame(spec = character(), aic = numeric(), cox_zph = numeric(), success = logical())
iiws_strat <- list()
comb_strata <- expand.grid(unique(data_mine_miss_restr_proc2$tipo_de_plan_2_mod),
unique(data_mine_miss_restr_proc2$time_interval3))
for (xx in seq_along(1:nrow(comb_strata))) {
i<- comb_strata[xx,"Var1"]
j<- comb_strata[xx,"Var2"]
data_mine_miss_proc2_subset<-
subset(data_mine_miss_restr_proc2, subset= tipo_de_plan_2_mod==i & time_interval3==j)
maxfu_df_subset <- subset(maxfu_restr_df, subset= maxfu.id %in% unique(data_mine_miss_proc2_subset$id))
iiw_model_strat<-
IrregLong::iiw.weights(Surv(time.lag,time,event)~
cluster(id)+ #If a frailty model is used, the cluster(id) term should appear before other covariates
tr_outcome.lag+
less_90d_tr1.lag+
log_dias_treat_imp_sin_na.lag +
comp_bpsc_y3_severe.lag+
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= data_mine_miss_proc2_subset,
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= maxfu_df_subset,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_oh", "susinidum_coc", "susinidum_pbc", "susinidum_mar", "psycom_dum_with", "psycom_dum_study_rec2", "freq_cons_dum_5day", "cond_oc_dum_3unemp", "cond_oc_dum_2inact", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "tr_outcome","log_dias_treat_imp_sin_na", "less_90d_tr1","comp_bpsc_y3_severe", "policonsumo2"),
lagfirst= c(2.95082,0,4.499811/2,0,0,0), #90/30.5 4.499811 es 90 días
first= T
)
iiws_strat[[xx]] <- list(
model = iiw_model_strat,
comb = paste0("Modality: ",i,"\nTime interval: ", j,"."),
cox_iiw = summary(iiw_model_strat$m),
iiw = summary(iiw_model_strat$iiw)
)
print(paste0("Modality: ",i,"\nTime interval: ", j))
print(summary(iiw_model_strat$m))
print(summary(iiw_model_strat$iiw))
}[1] "Modality: WO residential\nTime interval: [0,88]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 2106, number of events= 2106
(1417 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.19320 0.82432 0.05878 0.05420 -3.56
less_90d_tr1.lag -0.20058 0.81825 0.06876 0.08209 -2.44
log_dias_treat_imp_sin_na.lag -0.14656 0.86367 0.01558 0.02449 -5.98
comp_bpsc_y3_severe.lag -0.26031 0.77081 0.05617 0.06067 -4.29
policonsumo2.lag -0.43585 0.64671 0.06064 0.06111 -7.13
edad_al_ing_1 0.11002 1.11630 0.01083 0.00926 11.89
ano_nac_corr 0.12040 1.12795 0.01059 0.00897 13.42
susinidum_oh 0.21004 1.23372 0.16479 0.13855 1.52
susinidum_coc 0.27717 1.31939 0.18873 0.15659 1.77
susinidum_pbc 0.21194 1.23607 0.17095 0.14735 1.44
susinidum_mar 0.26304 1.30088 0.16570 0.14167 1.86
psycom_dum_with 0.07609 1.07906 0.06264 0.05653 1.35
psycom_dum_study -0.04540 0.95561 0.08076 0.07887 -0.58
freq_cons_dum_5day -0.00615 0.99387 0.05188 0.04702 -0.13
cond_oc_dum_2inact 0.02411 1.02441 0.08128 0.08206 0.29
cond_oc_dum_3unemp 0.11695 1.12406 0.08213 0.08238 1.42
susprindum_oh 0.10001 1.10518 0.15112 0.13956 0.72
susprindum_coc -0.00818 0.99185 0.15385 0.14136 -0.06
susprindum_pbc 0.08569 1.08947 0.14360 0.13471 0.64
susprindum_mar -0.03637 0.96429 0.21621 0.18971 -0.19
Pr(>|z|)
tr_outcome.lag 0.00036 ***
less_90d_tr1.lag 0.01455 *
log_dias_treat_imp_sin_na.lag 2.2e-09 ***
comp_bpsc_y3_severe.lag 1.8e-05 ***
policonsumo2.lag 9.9e-13 ***
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 0.12952
susinidum_coc 0.07673 .
susinidum_pbc 0.15034
susinidum_mar 0.06334 .
psycom_dum_with 0.17828
psycom_dum_study 0.56488
freq_cons_dum_5day 0.89601
cond_oc_dum_2inact 0.76885
cond_oc_dum_3unemp 0.15573
susprindum_oh 0.47362
susprindum_coc 0.95386
susprindum_pbc 0.52469
susprindum_mar 0.84798
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.824 1.213 0.741 0.917
less_90d_tr1.lag 0.818 1.222 0.697 0.961
log_dias_treat_imp_sin_na.lag 0.864 1.158 0.823 0.906
comp_bpsc_y3_severe.lag 0.771 1.297 0.684 0.868
policonsumo2.lag 0.647 1.546 0.574 0.729
edad_al_ing_1 1.116 0.896 1.096 1.137
ano_nac_corr 1.128 0.887 1.108 1.148
susinidum_oh 1.234 0.811 0.940 1.619
susinidum_coc 1.319 0.758 0.971 1.793
susinidum_pbc 1.236 0.809 0.926 1.650
susinidum_mar 1.301 0.769 0.985 1.717
psycom_dum_with 1.079 0.927 0.966 1.205
psycom_dum_study 0.956 1.046 0.819 1.115
freq_cons_dum_5day 0.994 1.006 0.906 1.090
cond_oc_dum_2inact 1.024 0.976 0.872 1.203
cond_oc_dum_3unemp 1.124 0.890 0.956 1.321
susprindum_oh 1.105 0.905 0.841 1.453
susprindum_coc 0.992 1.008 0.752 1.308
susprindum_pbc 1.089 0.918 0.837 1.419
susprindum_mar 0.964 1.037 0.665 1.399
Concordance= 0.653 (se = 0.005 )
Likelihood ratio test= 529 on 20 df, p=<2e-16
Wald test = 611 on 20 df, p=<2e-16
Score (logrank) test = 555 on 20 df, p=<2e-16, Robust = 539 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.20 0.58 1.09 1.28 1.87 4.98
[1] "Modality: GP intensive ambulatory\nTime interval: [0,88]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 9973, number of events= 9973
(6233 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.41578 0.65982 0.02667 0.03341 -12.44
less_90d_tr1.lag -0.21477 0.80673 0.03249 0.04867 -4.41
log_dias_treat_imp_sin_na.lag -0.21470 0.80678 0.00596 0.02474 -8.68
comp_bpsc_y3_severe.lag -0.17022 0.84348 0.02524 0.02689 -6.33
policonsumo2.lag -0.50518 0.60340 0.02670 0.03880 -13.02
edad_al_ing_1 0.12951 1.13827 0.00497 0.00468 27.69
ano_nac_corr 0.13571 1.14535 0.00490 0.00455 29.85
susinidum_oh -0.02056 0.97965 0.06953 0.06257 -0.33
susinidum_coc 0.01638 1.01652 0.08325 0.07644 0.21
susinidum_pbc 0.06987 1.07237 0.07878 0.07419 0.94
susinidum_mar 0.08715 1.09106 0.07084 0.06432 1.35
psycom_dum_with 0.03784 1.03857 0.02222 0.02392 1.58
psycom_dum_study 0.04332 1.04427 0.03479 0.04956 0.87
freq_cons_dum_5day 0.07887 1.08207 0.02063 0.02328 3.39
cond_oc_dum_2inact -0.00544 0.99458 0.03099 0.03969 -0.14
cond_oc_dum_3unemp 0.03677 1.03746 0.02222 0.02389 1.54
susprindum_oh -0.03507 0.96554 0.09009 0.10089 -0.35
susprindum_coc 0.06420 1.06631 0.09125 0.10167 0.63
susprindum_pbc 0.11017 1.11647 0.08896 0.10047 1.10
susprindum_mar 0.00978 1.00983 0.09767 0.11394 0.09
Pr(>|z|)
tr_outcome.lag < 2e-16 ***
less_90d_tr1.lag 1.0e-05 ***
log_dias_treat_imp_sin_na.lag < 2e-16 ***
comp_bpsc_y3_severe.lag 2.4e-10 ***
policonsumo2.lag < 2e-16 ***
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 0.7424
susinidum_coc 0.8303
susinidum_pbc 0.3463
susinidum_mar 0.1755
psycom_dum_with 0.1136
psycom_dum_study 0.3820
freq_cons_dum_5day 0.0007 ***
cond_oc_dum_2inact 0.8910
cond_oc_dum_3unemp 0.1238
susprindum_oh 0.7281
susprindum_coc 0.5277
susprindum_pbc 0.2728
susprindum_mar 0.9316
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.660 1.516 0.618 0.704
less_90d_tr1.lag 0.807 1.240 0.733 0.887
log_dias_treat_imp_sin_na.lag 0.807 1.239 0.769 0.847
comp_bpsc_y3_severe.lag 0.843 1.186 0.800 0.889
policonsumo2.lag 0.603 1.657 0.559 0.651
edad_al_ing_1 1.138 0.879 1.128 1.149
ano_nac_corr 1.145 0.873 1.135 1.156
susinidum_oh 0.980 1.021 0.867 1.107
susinidum_coc 1.017 0.984 0.875 1.181
susinidum_pbc 1.072 0.933 0.927 1.240
susinidum_mar 1.091 0.917 0.962 1.238
psycom_dum_with 1.039 0.963 0.991 1.088
psycom_dum_study 1.044 0.958 0.948 1.151
freq_cons_dum_5day 1.082 0.924 1.034 1.133
cond_oc_dum_2inact 0.995 1.005 0.920 1.075
cond_oc_dum_3unemp 1.037 0.964 0.990 1.087
susprindum_oh 0.966 1.036 0.792 1.177
susprindum_coc 1.066 0.938 0.874 1.301
susprindum_pbc 1.116 0.896 0.917 1.359
susprindum_mar 1.010 0.990 0.808 1.263
Concordance= 0.688 (se = 0.002 )
Likelihood ratio test= 3666 on 20 df, p=<2e-16
Wald test = 3057 on 20 df, p=<2e-16
Score (logrank) test = 4025 on 20 df, p=<2e-16, Robust = 3204 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.19 0.62 1.39 1.85 2.82 9.29
[1] "Modality: basic ambulatory\nTime interval: [0,88]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 8836, number of events= 8836
(5238 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.49853 0.60743 0.02786 0.04278 -11.65
less_90d_tr1.lag -0.29625 0.74360 0.03653 0.05216 -5.68
log_dias_treat_imp_sin_na.lag -0.26775 0.76510 0.00666 0.03198 -8.37
comp_bpsc_y3_severe.lag 0.03051 1.03098 0.03630 0.03660 0.83
policonsumo2.lag -0.50815 0.60161 0.02767 0.04248 -11.96
edad_al_ing_1 0.12296 1.13084 0.00541 0.00517 23.78
ano_nac_corr 0.13106 1.14004 0.00535 0.00523 25.04
susinidum_oh 0.03928 1.04007 0.07640 0.06470 0.61
susinidum_coc 0.10149 1.10682 0.09014 0.08332 1.22
susinidum_pbc 0.16419 1.17844 0.08742 0.07761 2.12
susinidum_mar 0.13452 1.14399 0.07752 0.06679 2.01
psycom_dum_with -0.00807 0.99196 0.02372 0.02648 -0.30
psycom_dum_study 0.17443 1.19057 0.03412 0.04334 4.03
freq_cons_dum_5day 0.00245 1.00245 0.02458 0.02389 0.10
cond_oc_dum_2inact 0.03825 1.03899 0.03378 0.03151 1.21
cond_oc_dum_3unemp 0.06070 1.06258 0.02600 0.02462 2.47
susprindum_oh 0.03566 1.03631 0.10103 0.07942 0.45
susprindum_coc 0.05788 1.05959 0.10247 0.08154 0.71
susprindum_pbc 0.13649 1.14625 0.10072 0.08066 1.69
susprindum_mar 0.10793 1.11397 0.10699 0.08789 1.23
Pr(>|z|)
tr_outcome.lag < 2e-16 ***
less_90d_tr1.lag 1.4e-08 ***
log_dias_treat_imp_sin_na.lag < 2e-16 ***
comp_bpsc_y3_severe.lag 0.404
policonsumo2.lag < 2e-16 ***
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 0.544
susinidum_coc 0.223
susinidum_pbc 0.034 *
susinidum_mar 0.044 *
psycom_dum_with 0.760
psycom_dum_study 5.7e-05 ***
freq_cons_dum_5day 0.918
cond_oc_dum_2inact 0.225
cond_oc_dum_3unemp 0.014 *
susprindum_oh 0.653
susprindum_coc 0.478
susprindum_pbc 0.091 .
susprindum_mar 0.219
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.607 1.646 0.559 0.661
less_90d_tr1.lag 0.744 1.345 0.671 0.824
log_dias_treat_imp_sin_na.lag 0.765 1.307 0.719 0.815
comp_bpsc_y3_severe.lag 1.031 0.970 0.960 1.108
policonsumo2.lag 0.602 1.662 0.554 0.654
edad_al_ing_1 1.131 0.884 1.119 1.142
ano_nac_corr 1.140 0.877 1.128 1.152
susinidum_oh 1.040 0.961 0.916 1.181
susinidum_coc 1.107 0.903 0.940 1.303
susinidum_pbc 1.178 0.849 1.012 1.372
susinidum_mar 1.144 0.874 1.004 1.304
psycom_dum_with 0.992 1.008 0.942 1.045
psycom_dum_study 1.191 0.840 1.094 1.296
freq_cons_dum_5day 1.002 0.998 0.957 1.051
cond_oc_dum_2inact 1.039 0.962 0.977 1.105
cond_oc_dum_3unemp 1.063 0.941 1.013 1.115
susprindum_oh 1.036 0.965 0.887 1.211
susprindum_coc 1.060 0.944 0.903 1.243
susprindum_pbc 1.146 0.872 0.979 1.343
susprindum_mar 1.114 0.898 0.938 1.323
Concordance= 0.692 (se = 0.003 )
Likelihood ratio test= 3513 on 20 df, p=<2e-16
Wald test = 2853 on 20 df, p=<2e-16
Score (logrank) test = 3910 on 20 df, p=<2e-16, Robust = 3228 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.11 0.54 1.34 1.80 2.78 11.57
[1] "Modality: GP residential\nTime interval: [0,88]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 4436, number of events= 4436
(2815 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.16991 0.84374 0.03876 0.03958 -4.29
less_90d_tr1.lag -0.19179 0.82548 0.04678 0.05353 -3.58
log_dias_treat_imp_sin_na.lag -0.08727 0.91643 0.00829 0.01870 -4.67
comp_bpsc_y3_severe.lag -0.24899 0.77959 0.03621 0.04101 -6.07
policonsumo2.lag -0.60356 0.54686 0.03971 0.04967 -12.15
edad_al_ing_1 0.09289 1.09734 0.00765 0.00648 14.34
ano_nac_corr 0.09753 1.10245 0.00760 0.00638 15.28
susinidum_oh 0.11062 1.11697 0.10673 0.12937 0.86
susinidum_coc 0.17321 1.18912 0.13021 0.14678 1.18
susinidum_pbc 0.17984 1.19703 0.11357 0.13542 1.33
susinidum_mar 0.29456 1.34254 0.10690 0.12899 2.28
psycom_dum_with -0.04422 0.95675 0.03443 0.03172 -1.39
psycom_dum_study 0.01251 1.01259 0.05082 0.05272 0.24
freq_cons_dum_5day 0.08040 1.08372 0.03485 0.03422 2.35
cond_oc_dum_2inact 0.09636 1.10116 0.05613 0.05494 1.75
cond_oc_dum_3unemp 0.19515 1.21549 0.04615 0.04405 4.43
susprindum_oh -0.11367 0.89255 0.16054 0.14683 -0.77
susprindum_coc -0.02726 0.97311 0.16143 0.14574 -0.19
susprindum_pbc -0.04708 0.95401 0.15660 0.14174 -0.33
susprindum_mar -0.16196 0.85048 0.18174 0.18838 -0.86
Pr(>|z|)
tr_outcome.lag 1.8e-05 ***
less_90d_tr1.lag 0.00034 ***
log_dias_treat_imp_sin_na.lag 3.1e-06 ***
comp_bpsc_y3_severe.lag 1.3e-09 ***
policonsumo2.lag < 2e-16 ***
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 0.39250
susinidum_coc 0.23796
susinidum_pbc 0.18418
susinidum_mar 0.02239 *
psycom_dum_with 0.16332
psycom_dum_study 0.81244
freq_cons_dum_5day 0.01878 *
cond_oc_dum_2inact 0.07944 .
cond_oc_dum_3unemp 9.4e-06 ***
susprindum_oh 0.43884
susprindum_coc 0.85165
susprindum_pbc 0.73975
susprindum_mar 0.38994
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.844 1.185 0.781 0.912
less_90d_tr1.lag 0.825 1.211 0.743 0.917
log_dias_treat_imp_sin_na.lag 0.916 1.091 0.883 0.951
comp_bpsc_y3_severe.lag 0.780 1.283 0.719 0.845
policonsumo2.lag 0.547 1.829 0.496 0.603
edad_al_ing_1 1.097 0.911 1.084 1.111
ano_nac_corr 1.102 0.907 1.089 1.116
susinidum_oh 1.117 0.895 0.867 1.439
susinidum_coc 1.189 0.841 0.892 1.586
susinidum_pbc 1.197 0.835 0.918 1.561
susinidum_mar 1.343 0.745 1.043 1.729
psycom_dum_with 0.957 1.045 0.899 1.018
psycom_dum_study 1.013 0.988 0.913 1.123
freq_cons_dum_5day 1.084 0.923 1.013 1.159
cond_oc_dum_2inact 1.101 0.908 0.989 1.226
cond_oc_dum_3unemp 1.215 0.823 1.115 1.325
susprindum_oh 0.893 1.120 0.669 1.190
susprindum_coc 0.973 1.028 0.731 1.295
susprindum_pbc 0.954 1.048 0.723 1.260
susprindum_mar 0.850 1.176 0.588 1.230
Concordance= 0.648 (se = 0.004 )
Likelihood ratio test= 1017 on 20 df, p=<2e-16
Wald test = 1099 on 20 df, p=<2e-16
Score (logrank) test = 1052 on 20 df, p=<2e-16, Robust = 998 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.27 0.66 1.22 1.42 2.07 4.40
[1] "Modality: WO intensive ambulatory\nTime interval: [0,88]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1546, number of events= 1546
(908 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.1752 0.8393 0.0737 0.0693 -2.53
less_90d_tr1.lag -0.3547 0.7014 0.0956 0.0974 -3.64
log_dias_treat_imp_sin_na.lag -0.4430 0.6421 0.0288 0.0285 -15.52
comp_bpsc_y3_severe.lag -0.0625 0.9394 0.0658 0.0629 -0.99
policonsumo2.lag -0.1927 0.8247 0.0777 0.0715 -2.69
edad_al_ing_1 0.1324 1.1416 0.0130 0.0117 11.30
ano_nac_corr 0.1351 1.1447 0.0127 0.0113 11.96
susinidum_oh -0.1597 0.8524 0.1803 0.1983 -0.81
susinidum_coc 0.0881 1.0921 0.2123 0.2232 0.39
susinidum_pbc -0.0693 0.9331 0.1926 0.2043 -0.34
susinidum_mar 0.0105 1.0105 0.1817 0.2002 0.05
psycom_dum_with -0.0250 0.9753 0.0579 0.0539 -0.46
psycom_dum_study -0.0806 0.9225 0.0915 0.0945 -0.85
freq_cons_dum_5day 0.0368 1.0375 0.0530 0.0513 0.72
cond_oc_dum_2inact 0.1451 1.1561 0.0667 0.0637 2.28
cond_oc_dum_3unemp 0.0518 1.0531 0.0692 0.0694 0.75
susprindum_oh -0.2115 0.8094 0.1931 0.1562 -1.35
susprindum_coc -0.0335 0.9671 0.1974 0.1628 -0.21
susprindum_pbc -0.0755 0.9273 0.1908 0.1590 -0.47
susprindum_mar -0.2156 0.8060 0.2084 0.1865 -1.16
Pr(>|z|)
tr_outcome.lag 0.01154 *
less_90d_tr1.lag 0.00027 ***
log_dias_treat_imp_sin_na.lag < 2e-16 ***
comp_bpsc_y3_severe.lag 0.32051
policonsumo2.lag 0.00705 **
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 0.42062
susinidum_coc 0.69305
susinidum_pbc 0.73457
susinidum_mar 0.95833
psycom_dum_with 0.64313
psycom_dum_study 0.39373
freq_cons_dum_5day 0.47326
cond_oc_dum_2inact 0.02281 *
cond_oc_dum_3unemp 0.45561
susprindum_oh 0.17571
susprindum_coc 0.83715
susprindum_pbc 0.63501
susprindum_mar 0.24748
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.839 1.191 0.733 0.962
less_90d_tr1.lag 0.701 1.426 0.579 0.849
log_dias_treat_imp_sin_na.lag 0.642 1.557 0.607 0.679
comp_bpsc_y3_severe.lag 0.939 1.064 0.830 1.063
policonsumo2.lag 0.825 1.213 0.717 0.949
edad_al_ing_1 1.142 0.876 1.116 1.168
ano_nac_corr 1.145 0.874 1.120 1.170
susinidum_oh 0.852 1.173 0.578 1.257
susinidum_coc 1.092 0.916 0.705 1.691
susinidum_pbc 0.933 1.072 0.625 1.393
susinidum_mar 1.011 0.990 0.682 1.496
psycom_dum_with 0.975 1.025 0.878 1.084
psycom_dum_study 0.923 1.084 0.766 1.110
freq_cons_dum_5day 1.037 0.964 0.938 1.147
cond_oc_dum_2inact 1.156 0.865 1.020 1.310
cond_oc_dum_3unemp 1.053 0.950 0.919 1.206
susprindum_oh 0.809 1.236 0.596 1.099
susprindum_coc 0.967 1.034 0.703 1.331
susprindum_pbc 0.927 1.078 0.679 1.266
susprindum_mar 0.806 1.241 0.559 1.162
Concordance= 0.692 (se = 0.006 )
Likelihood ratio test= 626 on 20 df, p=<2e-16
Wald test = 626 on 20 df, p=<2e-16
Score (logrank) test = 659 on 20 df, p=<2e-16, Robust = 544 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.21 0.54 1.46 1.93 2.89 13.67
[1] "Modality: WO residential\nTime interval: (93,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 34, number of events= 34
(32 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -1.44e+01 5.81e-07 7.40e+00 6.96e+00 -2.06
less_90d_tr1.lag NA NA 0.00e+00 0.00e+00 NA
log_dias_treat_imp_sin_na.lag 4.42e+00 8.31e+01 2.12e+00 1.90e+00 2.32
comp_bpsc_y3_severe.lag NA NA 0.00e+00 0.00e+00 NA
policonsumo2.lag NA NA 0.00e+00 0.00e+00 NA
edad_al_ing_1 2.49e+00 1.20e+01 7.01e-01 8.29e-01 3.00
ano_nac_corr 2.37e+00 1.07e+01 6.78e-01 8.04e-01 2.95
susinidum_oh -4.33e+00 1.32e-02 1.64e+00 1.38e+00 -3.14
susinidum_coc NA NA 0.00e+00 0.00e+00 NA
susinidum_pbc -4.56e+00 1.05e-02 1.69e+00 1.54e+00 -2.96
susinidum_mar -4.08e+00 1.69e-02 1.53e+00 1.07e+00 -3.81
psycom_dum_with 3.15e-01 1.37e+00 6.25e-01 5.93e-01 0.53
psycom_dum_study 1.83e+00 6.23e+00 1.23e+00 1.12e+00 1.63
freq_cons_dum_5day -2.80e-01 7.56e-01 5.60e-01 6.16e-01 -0.45
cond_oc_dum_2inact 6.22e-01 1.86e+00 1.24e+00 1.77e+00 0.35
cond_oc_dum_3unemp 1.16e+00 3.18e+00 1.15e+00 1.69e+00 0.68
susprindum_oh -1.67e+00 1.88e-01 1.93e+00 1.73e+00 -0.97
susprindum_coc -5.05e-01 6.04e-01 1.46e+00 1.09e+00 -0.46
susprindum_pbc -5.15e-01 5.98e-01 1.21e+00 7.97e-01 -0.65
susprindum_mar NA NA 0.00e+00 0.00e+00 NA
Pr(>|z|)
tr_outcome.lag 0.03905 *
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag 0.02013 *
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 0.00272 **
ano_nac_corr 0.00313 **
susinidum_oh 0.00172 **
susinidum_coc NA
susinidum_pbc 0.00310 **
susinidum_mar 0.00014 ***
psycom_dum_with 0.59557
psycom_dum_study 0.10206
freq_cons_dum_5day 0.65023
cond_oc_dum_2inact 0.72545
cond_oc_dum_3unemp 0.49430
susprindum_oh 0.33431
susprindum_coc 0.64295
susprindum_pbc 0.51798
susprindum_mar NA
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 5.81e-07 1.72e+06 6.94e-13 0.486
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag 8.31e+01 1.20e-02 2.00e+00 3452.547
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 1.20e+01 8.33e-02 2.36e+00 60.966
ano_nac_corr 1.07e+01 9.31e-02 2.22e+00 51.931
susinidum_oh 1.32e-02 7.59e+01 8.79e-04 0.197
susinidum_coc NA NA NA NA
susinidum_pbc 1.05e-02 9.52e+01 5.13e-04 0.215
susinidum_mar 1.69e-02 5.92e+01 2.07e-03 0.138
psycom_dum_with 1.37e+00 7.30e-01 4.28e-01 4.384
psycom_dum_study 6.23e+00 1.61e-01 6.95e-01 55.791
freq_cons_dum_5day 7.56e-01 1.32e+00 2.26e-01 2.531
cond_oc_dum_2inact 1.86e+00 5.37e-01 5.78e-02 60.021
cond_oc_dum_3unemp 3.18e+00 3.15e-01 1.15e-01 87.653
susprindum_oh 1.88e-01 5.31e+00 6.37e-03 5.578
susprindum_coc 6.04e-01 1.66e+00 7.14e-02 5.101
susprindum_pbc 5.98e-01 1.67e+00 1.25e-01 2.847
susprindum_mar NA NA NA NA
Concordance= 0.777 (se = 0.035 )
Likelihood ratio test= 29.1 on 15 df, p=0.02
Wald test = 93.5 on 15 df, p=2e-13
Score (logrank) test = 36.1 on 15 df, p=0.002, Robust = 28.5 p=0.02
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
3 28 94 179 288 729
[1] "Modality: GP intensive ambulatory\nTime interval: (93,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 213, number of events= 213
(202 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.0698 0.9325 0.7845 0.3906 -0.18
less_90d_tr1.lag 4.6058 100.0624 1.5181 0.7605 6.06
log_dias_treat_imp_sin_na.lag -0.1856 0.8306 0.2664 0.1478 -1.26
comp_bpsc_y3_severe.lag 0.3339 1.3965 0.6787 0.3763 0.89
policonsumo2.lag -0.4480 0.6389 0.7853 0.4145 -1.08
edad_al_ing_1 0.7114 2.0369 0.1093 0.1073 6.63
ano_nac_corr 0.7070 2.0279 0.1107 0.1092 6.47
susinidum_oh -0.0482 0.9530 0.3646 0.3959 -0.12
susinidum_coc 0.0251 1.0254 0.5269 0.4897 0.05
susinidum_pbc 0.0582 1.0599 0.5221 0.6476 0.09
susinidum_mar -0.0926 0.9116 0.3800 0.4148 -0.22
psycom_dum_with -0.1165 0.8901 0.1637 0.1548 -0.75
psycom_dum_study -0.1574 0.8544 0.2574 0.2633 -0.60
freq_cons_dum_5day 0.2429 1.2749 0.1658 0.1569 1.55
cond_oc_dum_2inact -0.0213 0.9789 0.2682 0.2480 -0.09
cond_oc_dum_3unemp -0.0428 0.9581 0.1665 0.1641 -0.26
susprindum_oh -0.8775 0.4158 0.6748 0.5933 -1.48
susprindum_coc -0.9359 0.3922 0.6822 0.6089 -1.54
susprindum_pbc -0.5899 0.5544 0.6602 0.5883 -1.00
susprindum_mar -0.7845 0.4564 0.7115 0.6598 -1.19
Pr(>|z|)
tr_outcome.lag 0.86
less_90d_tr1.lag 1.4e-09 ***
log_dias_treat_imp_sin_na.lag 0.21
comp_bpsc_y3_severe.lag 0.37
policonsumo2.lag 0.28
edad_al_ing_1 3.3e-11 ***
ano_nac_corr 9.6e-11 ***
susinidum_oh 0.90
susinidum_coc 0.96
susinidum_pbc 0.93
susinidum_mar 0.82
psycom_dum_with 0.45
psycom_dum_study 0.55
freq_cons_dum_5day 0.12
cond_oc_dum_2inact 0.93
cond_oc_dum_3unemp 0.79
susprindum_oh 0.14
susprindum_coc 0.12
susprindum_pbc 0.32
susprindum_mar 0.23
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.933 1.07234 0.434 2.01
less_90d_tr1.lag 100.062 0.00999 22.539 444.24
log_dias_treat_imp_sin_na.lag 0.831 1.20397 0.622 1.11
comp_bpsc_y3_severe.lag 1.396 0.71610 0.668 2.92
policonsumo2.lag 0.639 1.56513 0.284 1.44
edad_al_ing_1 2.037 0.49094 1.651 2.51
ano_nac_corr 2.028 0.49313 1.637 2.51
susinidum_oh 0.953 1.04934 0.439 2.07
susinidum_coc 1.025 0.97525 0.393 2.68
susinidum_pbc 1.060 0.94347 0.298 3.77
susinidum_mar 0.912 1.09702 0.404 2.06
psycom_dum_with 0.890 1.12353 0.657 1.21
psycom_dum_study 0.854 1.17043 0.510 1.43
freq_cons_dum_5day 1.275 0.78436 0.937 1.73
cond_oc_dum_2inact 0.979 1.02157 0.602 1.59
cond_oc_dum_3unemp 0.958 1.04374 0.695 1.32
susprindum_oh 0.416 2.40490 0.130 1.33
susprindum_coc 0.392 2.54955 0.119 1.29
susprindum_pbc 0.554 1.80387 0.175 1.76
susprindum_mar 0.456 2.19123 0.125 1.66
Concordance= 0.65 (se = 0.021 )
Likelihood ratio test= 67.7 on 20 df, p=4e-07
Wald test = 141 on 20 df, p=<2e-16
Score (logrank) test = 79.4 on 20 df, p=5e-09, Robust = 61.6 p=4e-06
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.07 1.45 2.09 2.66 3.06 15.52
[1] "Modality: basic ambulatory\nTime interval: (93,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 200, number of events= 200
(189 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.06539 1.06758 0.89003 0.52807 0.12
less_90d_tr1.lag NA NA 0.00000 0.00000 NA
log_dias_treat_imp_sin_na.lag -0.80835 0.44559 0.69768 0.41145 -1.96
comp_bpsc_y3_severe.lag 0.68033 1.97453 0.84695 0.44081 1.54
policonsumo2.lag 2.09839 8.15300 1.98344 1.21808 1.72
edad_al_ing_1 0.76099 2.14038 0.11599 0.12604 6.04
ano_nac_corr 0.74979 2.11655 0.11568 0.12543 5.98
susinidum_oh 0.52839 1.69620 0.52634 0.39173 1.35
susinidum_coc 0.17927 1.19634 0.58825 0.42912 0.42
susinidum_pbc 0.63893 1.89445 0.57098 0.43414 1.47
susinidum_mar 0.72591 2.06661 0.53943 0.40960 1.77
psycom_dum_with 0.03583 1.03648 0.17578 0.16452 0.22
psycom_dum_study 0.03533 1.03596 0.21114 0.20776 0.17
freq_cons_dum_5day 0.26250 1.30018 0.16544 0.16890 1.55
cond_oc_dum_2inact 0.27583 1.31763 0.23875 0.28028 0.98
cond_oc_dum_3unemp 0.00744 1.00747 0.18384 0.16345 0.05
susprindum_oh -0.71035 0.49147 0.51982 0.46058 -1.54
susprindum_coc -0.74138 0.47646 0.52322 0.46904 -1.58
susprindum_pbc -0.64819 0.52299 0.50722 0.44972 -1.44
susprindum_mar -0.20680 0.81319 0.55075 0.53491 -0.39
Pr(>|z|)
tr_outcome.lag 0.901
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag 0.049 *
comp_bpsc_y3_severe.lag 0.123
policonsumo2.lag 0.085 .
edad_al_ing_1 1.6e-09 ***
ano_nac_corr 2.3e-09 ***
susinidum_oh 0.177
susinidum_coc 0.676
susinidum_pbc 0.141
susinidum_mar 0.076 .
psycom_dum_with 0.828
psycom_dum_study 0.865
freq_cons_dum_5day 0.120
cond_oc_dum_2inact 0.325
cond_oc_dum_3unemp 0.964
susprindum_oh 0.123
susprindum_coc 0.114
susprindum_pbc 0.149
susprindum_mar 0.699
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.068 0.937 0.379 3.005
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag 0.446 2.244 0.199 0.998
comp_bpsc_y3_severe.lag 1.975 0.506 0.832 4.685
policonsumo2.lag 8.153 0.123 0.749 88.746
edad_al_ing_1 2.140 0.467 1.672 2.740
ano_nac_corr 2.117 0.472 1.655 2.706
susinidum_oh 1.696 0.590 0.787 3.655
susinidum_coc 1.196 0.836 0.516 2.774
susinidum_pbc 1.894 0.528 0.809 4.436
susinidum_mar 2.067 0.484 0.926 4.612
psycom_dum_with 1.036 0.965 0.751 1.431
psycom_dum_study 1.036 0.965 0.689 1.557
freq_cons_dum_5day 1.300 0.769 0.934 1.810
cond_oc_dum_2inact 1.318 0.759 0.761 2.282
cond_oc_dum_3unemp 1.007 0.993 0.731 1.388
susprindum_oh 0.491 2.035 0.199 1.212
susprindum_coc 0.476 2.099 0.190 1.195
susprindum_pbc 0.523 1.912 0.217 1.263
susprindum_mar 0.813 1.230 0.285 2.320
Concordance= 0.664 (se = 0.022 )
Likelihood ratio test= 66.7 on 19 df, p=3e-07
Wald test = 96.4 on 19 df, p=2e-12
Score (logrank) test = 65.2 on 19 df, p=6e-07, Robust = 66.6 p=3e-07
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.16 0.47 0.76 1.07 1.31 7.33
[1] "Modality: GP residential\nTime interval: (93,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 87, number of events= 87
(80 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.06200 1.06396 1.61401 1.02888 0.06
less_90d_tr1.lag 1.11330 3.04439 1.21710 0.96928 1.15
log_dias_treat_imp_sin_na.lag 2.21808 9.18970 1.19508 1.12830 1.97
comp_bpsc_y3_severe.lag -2.81004 0.06020 2.41466 2.01758 -1.39
policonsumo2.lag -5.68931 0.00338 2.59786 2.19430 -2.59
edad_al_ing_1 1.01020 2.74615 0.21856 0.19540 5.17
ano_nac_corr 0.99800 2.71284 0.21846 0.19453 5.13
susinidum_oh -1.06298 0.34542 1.05591 0.33960 -3.13
susinidum_coc -1.10348 0.33172 1.21863 0.52419 -2.11
susinidum_pbc -0.97188 0.37837 1.08524 0.48477 -2.00
susinidum_mar -1.19098 0.30392 1.06093 0.32881 -3.62
psycom_dum_with 0.10268 1.10814 0.30116 0.27290 0.38
psycom_dum_study 0.39160 1.47935 0.41112 0.32489 1.21
freq_cons_dum_5day -0.76394 0.46583 0.29504 0.25756 -2.97
cond_oc_dum_2inact -0.33948 0.71214 0.50269 0.56349 -0.60
cond_oc_dum_3unemp -0.73718 0.47846 0.41746 0.52470 -1.40
susprindum_oh -3.54379 0.02890 1.29324 0.70176 -5.05
susprindum_coc -2.46160 0.08530 1.29604 0.79542 -3.09
susprindum_pbc -2.26402 0.10393 1.27244 0.75037 -3.02
susprindum_mar -1.64439 0.19313 1.51917 0.81467 -2.02
Pr(>|z|)
tr_outcome.lag 0.95195
less_90d_tr1.lag 0.25073
log_dias_treat_imp_sin_na.lag 0.04931 *
comp_bpsc_y3_severe.lag 0.16369
policonsumo2.lag 0.00952 **
edad_al_ing_1 2.3e-07 ***
ano_nac_corr 2.9e-07 ***
susinidum_oh 0.00175 **
susinidum_coc 0.03528 *
susinidum_pbc 0.04498 *
susinidum_mar 0.00029 ***
psycom_dum_with 0.70672
psycom_dum_study 0.22807
freq_cons_dum_5day 0.00302 **
cond_oc_dum_2inact 0.54687
cond_oc_dum_3unemp 0.16003
susprindum_oh 4.4e-07 ***
susprindum_coc 0.00197 **
susprindum_pbc 0.00255 **
susprindum_mar 0.04354 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.06396 0.940 1.42e-01 7.993
less_90d_tr1.lag 3.04439 0.328 4.55e-01 20.350
log_dias_treat_imp_sin_na.lag 9.18970 0.109 1.01e+00 83.890
comp_bpsc_y3_severe.lag 0.06020 16.611 1.15e-03 3.140
policonsumo2.lag 0.00338 295.691 4.59e-05 0.249
edad_al_ing_1 2.74615 0.364 1.87e+00 4.028
ano_nac_corr 2.71284 0.369 1.85e+00 3.972
susinidum_oh 0.34542 2.895 1.78e-01 0.672
susinidum_coc 0.33172 3.015 1.19e-01 0.927
susinidum_pbc 0.37837 2.643 1.46e-01 0.978
susinidum_mar 0.30392 3.290 1.60e-01 0.579
psycom_dum_with 1.10814 0.902 6.49e-01 1.892
psycom_dum_study 1.47935 0.676 7.83e-01 2.796
freq_cons_dum_5day 0.46583 2.147 2.81e-01 0.772
cond_oc_dum_2inact 0.71214 1.404 2.36e-01 2.149
cond_oc_dum_3unemp 0.47846 2.090 1.71e-01 1.338
susprindum_oh 0.02890 34.598 7.30e-03 0.114
susprindum_coc 0.08530 11.724 1.79e-02 0.406
susprindum_pbc 0.10393 9.622 2.39e-02 0.452
susprindum_mar 0.19313 5.178 3.91e-02 0.953
Concordance= 0.729 (se = 0.029 )
Likelihood ratio test= 53.7 on 20 df, p=6e-05
Wald test = 118 on 20 df, p=8e-16
Score (logrank) test = 49 on 20 df, p=3e-04, Robust = 43.5 p=0.002
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
17 68 123 305 261 3240
[1] "Modality: WO intensive ambulatory\nTime interval: (93,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 41, number of events= 41
(36 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 1.1304 3.0969 1.5680 0.9116 1.24
less_90d_tr1.lag NA NA 0.0000 0.0000 NA
log_dias_treat_imp_sin_na.lag 0.1992 1.2204 0.2918 0.2185 0.91
comp_bpsc_y3_severe.lag -3.0324 0.0482 1.7463 1.0916 -2.78
policonsumo2.lag -3.5939 0.0275 2.0412 1.5108 -2.38
edad_al_ing_1 0.4265 1.5319 0.3017 0.3236 1.32
ano_nac_corr 0.3897 1.4765 0.2918 0.3157 1.23
susinidum_oh -1.2588 0.2840 0.6121 0.7170 -1.76
susinidum_coc -3.3142 0.0364 1.3930 1.0273 -3.23
susinidum_pbc -1.3977 0.2472 0.7332 0.8635 -1.62
susinidum_mar NA NA 0.0000 0.0000 NA
psycom_dum_with 0.4333 1.5423 0.5645 0.5095 0.85
psycom_dum_study -0.3400 0.7118 0.6199 0.6590 -0.52
freq_cons_dum_5day 1.9088 6.7448 0.6268 0.6090 3.13
cond_oc_dum_2inact -1.7361 0.1762 0.9768 1.0968 -1.58
cond_oc_dum_3unemp -1.7825 0.1682 1.0476 1.0569 -1.69
susprindum_oh -1.8979 0.1499 1.5268 1.5030 -1.26
susprindum_coc -0.6726 0.5104 1.2308 0.7463 -0.90
susprindum_pbc -0.6693 0.5121 1.1493 0.7270 -0.92
susprindum_mar -1.7836 0.1680 1.4825 1.1096 -1.61
Pr(>|z|)
tr_outcome.lag 0.2149
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag 0.3620
comp_bpsc_y3_severe.lag 0.0055 **
policonsumo2.lag 0.0174 *
edad_al_ing_1 0.1875
ano_nac_corr 0.2171
susinidum_oh 0.0791 .
susinidum_coc 0.0013 **
susinidum_pbc 0.1055
susinidum_mar NA
psycom_dum_with 0.3951
psycom_dum_study 0.6059
freq_cons_dum_5day 0.0017 **
cond_oc_dum_2inact 0.1135
cond_oc_dum_3unemp 0.0917 .
susprindum_oh 0.2067
susprindum_coc 0.3675
susprindum_pbc 0.3572
susprindum_mar 0.1079
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 3.0969 0.323 0.51881 18.486
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag 1.2204 0.819 0.79528 1.873
comp_bpsc_y3_severe.lag 0.0482 20.748 0.00567 0.409
policonsumo2.lag 0.0275 36.376 0.00142 0.531
edad_al_ing_1 1.5319 0.653 0.81240 2.888
ano_nac_corr 1.4765 0.677 0.79523 2.741
susinidum_oh 0.2840 3.521 0.06966 1.158
susinidum_coc 0.0364 27.502 0.00485 0.272
susinidum_pbc 0.2472 4.046 0.04550 1.343
susinidum_mar NA NA NA NA
psycom_dum_with 1.5423 0.648 0.56817 4.187
psycom_dum_study 0.7118 1.405 0.19560 2.590
freq_cons_dum_5day 6.7448 0.148 2.04461 22.250
cond_oc_dum_2inact 0.1762 5.675 0.02053 1.512
cond_oc_dum_3unemp 0.1682 5.945 0.02119 1.335
susprindum_oh 0.1499 6.672 0.00788 2.852
susprindum_coc 0.5104 1.959 0.11820 2.204
susprindum_pbc 0.5121 1.953 0.12317 2.129
susprindum_mar 0.1680 5.951 0.01910 1.479
Concordance= 0.714 (se = 0.049 )
Likelihood ratio test= 23.5 on 18 df, p=0.2
Wald test = 191 on 18 df, p=<2e-16
Score (logrank) test = 26.4 on 18 df, p=0.09, Robust = 19.6 p=0.4
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
1 7 15 34 32 354
[1] "Modality: WO residential\nTime interval: (88,93]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 13, number of events= 13
(13 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag NA NA 0.00e+00 0.00e+00 NA
less_90d_tr1.lag NA NA 0.00e+00 0.00e+00 NA
log_dias_treat_imp_sin_na.lag NA NA 0.00e+00 0.00e+00 NA
comp_bpsc_y3_severe.lag NA NA 0.00e+00 0.00e+00 NA
policonsumo2.lag NA NA 0.00e+00 0.00e+00 NA
edad_al_ing_1 -5.69e+01 2.01e-25 5.83e+02 2.31e+00 -24.7
ano_nac_corr -3.70e+01 8.29e-17 2.42e+04 1.94e+00 -19.1
susinidum_oh 2.40e+02 1.77e+104 3.87e+05 7.01e+00 34.2
susinidum_coc NA NA 0.00e+00 0.00e+00 NA
susinidum_pbc 1.68e+02 9.89e+72 1.69e+05 5.26e+00 31.9
susinidum_mar NA NA 0.00e+00 0.00e+00 NA
psycom_dum_with -2.17e+01 3.76e-10 7.26e+04 1.38e+00 -15.7
psycom_dum_study -1.21e+02 1.81e-53 4.11e+05 5.48e+00 -22.2
freq_cons_dum_5day -6.09e+01 3.71e-27 2.42e+04 2.26e+00 -26.9
cond_oc_dum_2inact NA NA 0.00e+00 0.00e+00 NA
cond_oc_dum_3unemp NA NA 0.00e+00 0.00e+00 NA
susprindum_oh 9.63e+01 6.38e+41 1.21e+05 5.27e+00 18.2
susprindum_coc 1.21e+02 2.66e+52 7.26e+04 3.47e+00 34.8
susprindum_pbc 1.14e+02 3.44e+49 2.43e+04 4.27e+00 26.7
susprindum_mar NA NA 0.00e+00 0.00e+00 NA
Pr(>|z|)
tr_outcome.lag NA
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag NA
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 <2e-16 ***
ano_nac_corr <2e-16 ***
susinidum_oh <2e-16 ***
susinidum_coc NA
susinidum_pbc <2e-16 ***
susinidum_mar NA
psycom_dum_with <2e-16 ***
psycom_dum_study <2e-16 ***
freq_cons_dum_5day <2e-16 ***
cond_oc_dum_2inact NA
cond_oc_dum_3unemp NA
susprindum_oh <2e-16 ***
susprindum_coc <2e-16 ***
susprindum_pbc <2e-16 ***
susprindum_mar NA
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag NA NA NA NA
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag NA NA NA NA
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 2.01e-25 4.98e+24 2.19e-27 1.84e-23
ano_nac_corr 8.29e-17 1.21e+16 1.84e-18 3.73e-15
susinidum_oh 1.77e+104 5.64e-105 1.92e+98 1.64e+110
susinidum_coc NA NA NA NA
susinidum_pbc 9.89e+72 1.01e-73 3.30e+68 2.97e+77
susinidum_mar NA NA NA NA
psycom_dum_with 3.76e-10 2.66e+09 2.52e-11 5.61e-09
psycom_dum_study 1.81e-53 5.53e+52 3.91e-58 8.35e-49
freq_cons_dum_5day 3.71e-27 2.70e+26 4.43e-29 3.11e-25
cond_oc_dum_2inact NA NA NA NA
cond_oc_dum_3unemp NA NA NA NA
susprindum_oh 6.38e+41 1.57e-42 2.07e+37 1.97e+46
susprindum_coc 2.66e+52 3.76e-53 2.98e+49 2.37e+55
susprindum_pbc 3.44e+49 2.90e-50 7.99e+45 1.48e+53
susprindum_mar NA NA NA NA
Concordance= 1 (se = 0 )
Likelihood ratio test= 45.1 on 10 df, p=2e-06
Wald test = 1819 on 10 df, p=<2e-16
Score (logrank) test = 16.9 on 10 df, p=0.08, Robust = 12.6 p=0.2
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.00e+00 0.00e+00 0.00e+00 2.29e+68 0.00e+00 2.98e+69
[1] "Modality: GP intensive ambulatory\nTime interval: (88,93]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 74, number of events= 74
(74 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag NA NA 0.0000 0.0000 NA
less_90d_tr1.lag NA NA 0.0000 0.0000 NA
log_dias_treat_imp_sin_na.lag NA NA 0.0000 0.0000 NA
comp_bpsc_y3_severe.lag NA NA 0.0000 0.0000 NA
policonsumo2.lag NA NA 0.0000 0.0000 NA
edad_al_ing_1 0.0225 1.0228 0.1475 0.1616 0.14
ano_nac_corr 0.0460 1.0471 0.1395 0.1496 0.31
susinidum_oh -0.2291 0.7953 0.2985 0.2714 -0.84
susinidum_coc 0.8542 2.3495 0.5729 0.5836 1.46
susinidum_pbc -0.6890 0.5021 0.5586 0.6669 -1.03
susinidum_mar NA NA 0.0000 0.0000 NA
psycom_dum_with 0.3664 1.4425 0.3078 0.2991 1.22
psycom_dum_study 1.1222 3.0717 0.4851 0.4892 2.29
freq_cons_dum_5day 0.2846 1.3292 0.3179 0.2895 0.98
cond_oc_dum_2inact 0.1720 1.1877 0.4223 0.3660 0.47
cond_oc_dum_3unemp 0.4730 1.6048 0.3933 0.3670 1.29
susprindum_oh -0.0812 0.9220 0.6170 0.4583 -0.18
susprindum_coc -0.0655 0.9366 0.6616 0.5184 -0.13
susprindum_pbc -0.6944 0.4994 0.5289 0.3637 -1.91
susprindum_mar NA NA 0.0000 0.0000 NA
Pr(>|z|)
tr_outcome.lag NA
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag NA
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 0.889
ano_nac_corr 0.758
susinidum_oh 0.399
susinidum_coc 0.143
susinidum_pbc 0.302
susinidum_mar NA
psycom_dum_with 0.221
psycom_dum_study 0.022 *
freq_cons_dum_5day 0.326
cond_oc_dum_2inact 0.638
cond_oc_dum_3unemp 0.197
susprindum_oh 0.859
susprindum_coc 0.900
susprindum_pbc 0.056 .
susprindum_mar NA
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag NA NA NA NA
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag NA NA NA NA
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 1.023 0.978 0.745 1.40
ano_nac_corr 1.047 0.955 0.781 1.40
susinidum_oh 0.795 1.257 0.467 1.35
susinidum_coc 2.350 0.426 0.749 7.37
susinidum_pbc 0.502 1.992 0.136 1.86
susinidum_mar NA NA NA NA
psycom_dum_with 1.443 0.693 0.803 2.59
psycom_dum_study 3.072 0.326 1.177 8.01
freq_cons_dum_5day 1.329 0.752 0.754 2.34
cond_oc_dum_2inact 1.188 0.842 0.580 2.43
cond_oc_dum_3unemp 1.605 0.623 0.782 3.29
susprindum_oh 0.922 1.085 0.375 2.26
susprindum_coc 0.937 1.068 0.339 2.59
susprindum_pbc 0.499 2.002 0.245 1.02
susprindum_mar NA NA NA NA
Concordance= 0.634 (se = 0.034 )
Likelihood ratio test= 13.9 on 13 df, p=0.4
Wald test = 23.4 on 13 df, p=0.04
Score (logrank) test = 14.5 on 13 df, p=0.3, Robust = 16.6 p=0.2
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.198 0.615 0.804 0.936 1.277 2.215
[1] "Modality: basic ambulatory\nTime interval: (88,93]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 77, number of events= 77
(76 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag NA NA 0.00e+00 0.00e+00 NA
less_90d_tr1.lag NA NA 0.00e+00 0.00e+00 NA
log_dias_treat_imp_sin_na.lag -4.77e-01 6.21e-01 4.43e-01 2.14e-01 -2.23
comp_bpsc_y3_severe.lag NA NA 0.00e+00 0.00e+00 NA
policonsumo2.lag NA NA 0.00e+00 0.00e+00 NA
edad_al_ing_1 -8.12e-02 9.22e-01 1.57e-01 1.39e-01 -0.58
ano_nac_corr -1.72e-01 8.42e-01 1.59e-01 1.40e-01 -1.23
susinidum_oh -6.05e-01 5.46e-01 6.92e-01 4.73e-01 -1.28
susinidum_coc -8.26e-02 9.21e-01 1.05e+00 6.08e-01 -0.14
susinidum_pbc -1.81e+00 1.64e-01 8.99e-01 6.97e-01 -2.59
susinidum_mar -8.53e-01 4.26e-01 7.36e-01 5.27e-01 -1.62
psycom_dum_with -3.47e-01 7.07e-01 3.23e-01 2.73e-01 -1.27
psycom_dum_study -6.93e-01 5.00e-01 3.65e-01 3.41e-01 -2.03
freq_cons_dum_5day -2.08e-01 8.12e-01 3.02e-01 2.87e-01 -0.73
cond_oc_dum_2inact 7.39e-01 2.09e+00 4.50e-01 4.40e-01 1.68
cond_oc_dum_3unemp 4.68e-01 1.60e+00 3.39e-01 3.44e-01 1.36
susprindum_oh -2.10e+01 7.35e-10 3.08e+03 6.52e-01 -32.26
susprindum_coc -2.08e+01 9.01e-10 3.08e+03 5.09e-01 -40.94
susprindum_pbc -1.95e+01 3.43e-09 3.08e+03 5.67e-01 -34.39
susprindum_mar -2.06e+01 1.17e-09 3.08e+03 6.89e-01 -29.86
Pr(>|z|)
tr_outcome.lag NA
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag 0.0258 *
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 0.5597
ano_nac_corr 0.2186
susinidum_oh 0.2008
susinidum_coc 0.8919
susinidum_pbc 0.0095 **
susinidum_mar 0.1056
psycom_dum_with 0.2028
psycom_dum_study 0.0423 *
freq_cons_dum_5day 0.4680
cond_oc_dum_2inact 0.0926 .
cond_oc_dum_3unemp 0.1732
susprindum_oh <2e-16 ***
susprindum_coc <2e-16 ***
susprindum_pbc <2e-16 ***
susprindum_mar <2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag NA NA NA NA
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag 6.21e-01 1.61e+00 4.08e-01 9.44e-01
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 9.22e-01 1.08e+00 7.02e-01 1.21e+00
ano_nac_corr 8.42e-01 1.19e+00 6.41e-01 1.11e+00
susinidum_oh 5.46e-01 1.83e+00 2.16e-01 1.38e+00
susinidum_coc 9.21e-01 1.09e+00 2.80e-01 3.03e+00
susinidum_pbc 1.64e-01 6.11e+00 4.17e-02 6.43e-01
susinidum_mar 4.26e-01 2.35e+00 1.52e-01 1.20e+00
psycom_dum_with 7.07e-01 1.41e+00 4.14e-01 1.21e+00
psycom_dum_study 5.00e-01 2.00e+00 2.56e-01 9.76e-01
freq_cons_dum_5day 8.12e-01 1.23e+00 4.63e-01 1.42e+00
cond_oc_dum_2inact 2.09e+00 4.77e-01 8.85e-01 4.96e+00
cond_oc_dum_3unemp 1.60e+00 6.26e-01 8.14e-01 3.13e+00
susprindum_oh 7.35e-10 1.36e+09 2.05e-10 2.64e-09
susprindum_coc 9.01e-10 1.11e+09 3.33e-10 2.44e-09
susprindum_pbc 3.43e-09 2.91e+08 1.13e-09 1.04e-08
susprindum_mar 1.17e-09 8.58e+08 3.02e-10 4.50e-09
Concordance= 0.731 (se = 0.03 )
Likelihood ratio test= 44.5 on 16 df, p=2e-04
Wald test = 16181 on 16 df, p=<2e-16
Score (logrank) test = 111 on 16 df, p=2e-16, Robust = 37.4 p=0.002
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.00e+00 7.19e+08 1.20e+09 1.84e+09 2.58e+09 7.16e+09
[1] "Modality: GP residential\nTime interval: (88,93]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 44, number of events= 44
(44 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag NA NA 0.0000 0.0000 NA
less_90d_tr1.lag NA NA 0.0000 0.0000 NA
log_dias_treat_imp_sin_na.lag NA NA 0.0000 0.0000 NA
comp_bpsc_y3_severe.lag NA NA 0.0000 0.0000 NA
policonsumo2.lag NA NA 0.0000 0.0000 NA
edad_al_ing_1 -0.0941 0.9102 0.2559 0.2579 -0.36
ano_nac_corr -0.1054 0.9000 0.2603 0.2672 -0.39
susinidum_oh -1.3805 0.2515 1.1947 0.7912 -1.74
susinidum_coc -2.3484 0.0955 1.7688 1.0683 -2.20
susinidum_pbc -1.7372 0.1760 1.2633 0.9136 -1.90
susinidum_mar -1.7400 0.1755 1.2540 0.9334 -1.86
psycom_dum_with -0.5855 0.5568 0.4569 0.4924 -1.19
psycom_dum_study -0.5570 0.5729 0.7513 0.6086 -0.92
freq_cons_dum_5day 0.2100 1.2337 0.4609 0.4900 0.43
cond_oc_dum_2inact 1.6085 4.9952 0.8868 0.6243 2.58
cond_oc_dum_3unemp 0.6958 2.0053 0.6945 0.7280 0.96
susprindum_oh -3.2052 0.0406 1.4883 1.2502 -2.56
susprindum_coc -2.1759 0.1135 1.4852 1.1353 -1.92
susprindum_pbc -2.7616 0.0632 1.2925 0.9600 -2.88
susprindum_mar NA NA 0.0000 0.0000 NA
Pr(>|z|)
tr_outcome.lag NA
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag NA
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 0.715
ano_nac_corr 0.693
susinidum_oh 0.081 .
susinidum_coc 0.028 *
susinidum_pbc 0.057 .
susinidum_mar 0.062 .
psycom_dum_with 0.234
psycom_dum_study 0.360
freq_cons_dum_5day 0.668
cond_oc_dum_2inact 0.010 **
cond_oc_dum_3unemp 0.339
susprindum_oh 0.010 *
susprindum_coc 0.055 .
susprindum_pbc 0.004 **
susprindum_mar NA
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag NA NA NA NA
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag NA NA NA NA
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 0.9102 1.099 0.54904 1.509
ano_nac_corr 0.9000 1.111 0.53305 1.520
susinidum_oh 0.2515 3.977 0.05333 1.186
susinidum_coc 0.0955 10.469 0.01177 0.775
susinidum_pbc 0.1760 5.682 0.02937 1.055
susinidum_mar 0.1755 5.697 0.02817 1.094
psycom_dum_with 0.5568 1.796 0.21212 1.462
psycom_dum_study 0.5729 1.745 0.17380 1.889
freq_cons_dum_5day 1.2337 0.811 0.47219 3.223
cond_oc_dum_2inact 4.9952 0.200 1.46948 16.980
cond_oc_dum_3unemp 2.0053 0.499 0.48140 8.353
susprindum_oh 0.0406 24.660 0.00350 0.470
susprindum_coc 0.1135 8.810 0.01227 1.050
susprindum_pbc 0.0632 15.825 0.00963 0.415
susprindum_mar NA NA NA NA
Concordance= 0.62 (se = 0.051 )
Likelihood ratio test= 12.3 on 14 df, p=0.6
Wald test = 38.5 on 14 df, p=4e-04
Score (logrank) test = 20 on 14 df, p=0.1, Robust = 11.2 p=0.7
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
2.6 34.2 48.9 52.1 70.7 99.9
[1] "Modality: WO intensive ambulatory\nTime interval: (88,93]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 10, number of events= 10
(10 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag NA NA 0.00e+00 0.00e+00 NA
less_90d_tr1.lag NA NA 0.00e+00 0.00e+00 NA
log_dias_treat_imp_sin_na.lag NA NA 0.00e+00 0.00e+00 NA
comp_bpsc_y3_severe.lag NA NA 0.00e+00 0.00e+00 NA
policonsumo2.lag NA NA 0.00e+00 0.00e+00 NA
edad_al_ing_1 -5.34e+01 6.69e-24 6.64e+03 1.45e+00 -36.92
ano_nac_corr -4.99e+01 2.05e-22 5.79e+03 1.40e+00 -35.58
susinidum_oh 5.70e+00 2.99e+02 1.73e+03 1.07e+00 5.31
susinidum_coc NA NA 0.00e+00 0.00e+00 NA
susinidum_pbc 5.46e+01 5.41e+23 1.65e+04 3.47e+00 15.75
susinidum_mar NA NA 0.00e+00 0.00e+00 NA
psycom_dum_with 2.74e+01 8.14e+11 4.34e+03 1.45e+00 18.95
psycom_dum_study NA NA 0.00e+00 0.00e+00 NA
freq_cons_dum_5day -1.48e+02 6.27e-65 1.42e+04 3.58e+00 -41.25
cond_oc_dum_2inact -4.10e+01 1.49e-18 8.48e+03 2.02e+00 -20.30
cond_oc_dum_3unemp NA NA 0.00e+00 0.00e+00 NA
susprindum_oh 7.39e+01 1.23e+32 6.98e+03 3.27e+00 22.57
susprindum_coc 1.39e+02 1.89e+60 6.56e+03 3.67e+00 37.84
susprindum_pbc NA NA 0.00e+00 0.00e+00 NA
susprindum_mar NA NA 0.00e+00 0.00e+00 NA
Pr(>|z|)
tr_outcome.lag NA
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag NA
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 1.1e-07 ***
susinidum_coc NA
susinidum_pbc < 2e-16 ***
susinidum_mar NA
psycom_dum_with < 2e-16 ***
psycom_dum_study NA
freq_cons_dum_5day < 2e-16 ***
cond_oc_dum_2inact < 2e-16 ***
cond_oc_dum_3unemp NA
susprindum_oh < 2e-16 ***
susprindum_coc < 2e-16 ***
susprindum_pbc NA
susprindum_mar NA
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag NA NA NA NA
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag NA NA NA NA
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 6.69e-24 1.50e+23 3.94e-25 1.14e-22
ano_nac_corr 2.05e-22 4.87e+21 1.31e-23 3.21e-21
susinidum_oh 2.99e+02 3.34e-03 3.64e+01 2.45e+03
susinidum_coc NA NA NA NA
susinidum_pbc 5.41e+23 1.85e-24 6.03e+20 4.86e+26
susinidum_mar NA NA NA NA
psycom_dum_with 8.14e+11 1.23e-12 4.77e+10 1.39e+13
psycom_dum_study NA NA NA NA
freq_cons_dum_5day 6.27e-65 1.59e+64 5.58e-68 7.05e-62
cond_oc_dum_2inact 1.49e-18 6.70e+17 2.84e-20 7.85e-17
cond_oc_dum_3unemp NA NA NA NA
susprindum_oh 1.23e+32 8.13e-33 2.01e+29 7.53e+34
susprindum_coc 1.89e+60 5.28e-61 1.43e+57 2.51e+63
susprindum_pbc NA NA NA NA
susprindum_mar NA NA NA NA
Concordance= 1 (se = 0 )
Likelihood ratio test= 30.2 on 9 df, p=4e-04
Wald test = 1857 on 9 df, p=<2e-16
Score (logrank) test = 21.2 on 9 df, p=0.01, Robust = 9.99 p=0.4
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.00e+00 0.00e+00 8.00e+01 1.71e+24 2.95e+13 1.71e+25 Code
invisible("This model had some problems, not stable (low n in strata)")
invisible("Did it later, at 2024-05-16")
iiws_strat1 <- list()
comb_strata1 <- expand.grid(unique(data_mine_miss_restr_proc2$tipo_de_plan_2_mod),
unique(data_mine_miss_restr_proc2$time_interval3))
for (xx in seq_along(1:nrow(comb_strata1))) {
i<- comb_strata1[xx,"Var1"]
j<- comb_strata1[xx,"Var2"]
data_mine_miss_proc2_subset<-
subset(data_mine_miss_restr_proc2, subset= tipo_de_plan_2_mod==i & time_interval3==j)
maxfu_df_subset <- subset(maxfu_restr_df, subset= maxfu.id %in% unique(data_mine_miss_proc2_subset$id))
iiw_model_strat1<-
IrregLong::iiw.weights(Surv(time.lag,time,event)~
cluster(id)+ #If a frailty model is used, the cluster(id) term should appear before other covariates
tr_outcome.lag+
less_90d_tr1.lag+
log_dias_treat_imp_sin_na.lag +
comp_bpsc_y3_severe.lag+
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= data_mine_miss_proc2_subset,
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= maxfu_df_subset,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_oh", "susinidum_coc", "susinidum_pbc", "susinidum_mar", "psycom_dum_with", "psycom_dum_study_rec2", "freq_cons_dum_5day", "cond_oc_dum_3unemp", "cond_oc_dum_2inact", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "tr_outcome","log_dias_treat_imp_sin_na", "less_90d_tr1","comp_bpsc_y3_severe", "policonsumo2"),
lagfirst= c(2.95082,1,4.499811,1,1,1), #90/30.5 4.499811 es 90 días
first= T
)
iiws_strat1[[xx]] <- list(
model = iiw_model_strat1,
comb = paste0("Modality: ",i,"\nTime interval: ", j,"."),
cox_iiw = summary(iiw_model_strat1$m),
iiw = summary(iiw_model_strat1$iiw)
)
print(paste0("Modality: ",i,"\nTime interval: ", j))
print(summary(iiw_model_strat1$m))
print(summary(iiw_model_strat1$iiw))
}[1] "Modality: WO residential\nTime interval: [0,88]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 2106, number of events= 2106
(1417 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.20536 1.22796 0.06616 0.05418 3.79
less_90d_tr1.lag 0.80271 2.23158 0.07067 0.07676 10.46
log_dias_treat_imp_sin_na.lag 0.17616 1.19262 0.03140 0.03679 4.79
comp_bpsc_y3_severe.lag 0.17454 1.19070 0.06655 0.07093 2.46
policonsumo2.lag -0.04576 0.95527 0.07395 0.05745 -0.80
edad_al_ing_1 0.11274 1.11935 0.01083 0.00973 11.59
ano_nac_corr 0.11840 1.12569 0.01057 0.00935 12.66
susinidum_oh 0.03767 1.03839 0.16591 0.14173 0.27
susinidum_coc 0.08672 1.09060 0.18995 0.16127 0.54
susinidum_pbc 0.02071 1.02093 0.17193 0.15333 0.14
susinidum_mar 0.08361 1.08721 0.16692 0.14457 0.58
psycom_dum_with -0.02497 0.97534 0.06301 0.05548 -0.45
psycom_dum_study -0.39981 0.67045 0.07750 0.07329 -5.45
freq_cons_dum_5day -0.04344 0.95749 0.05145 0.05074 -0.86
cond_oc_dum_2inact 0.03819 1.03893 0.08141 0.08997 0.42
cond_oc_dum_3unemp 0.08288 1.08641 0.08212 0.09092 0.91
susprindum_oh -0.03090 0.96957 0.15198 0.14127 -0.22
susprindum_coc -0.10130 0.90367 0.15487 0.14385 -0.70
susprindum_pbc -0.06907 0.93326 0.14478 0.13566 -0.51
susprindum_mar -0.25578 0.77432 0.21774 0.20267 -1.26
Pr(>|z|)
tr_outcome.lag 0.00015 ***
less_90d_tr1.lag < 2e-16 ***
log_dias_treat_imp_sin_na.lag 1.7e-06 ***
comp_bpsc_y3_severe.lag 0.01386 *
policonsumo2.lag 0.42565
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 0.79041
susinidum_coc 0.59075
susinidum_pbc 0.89254
susinidum_mar 0.56304
psycom_dum_with 0.65264
psycom_dum_study 4.9e-08 ***
freq_cons_dum_5day 0.39189
cond_oc_dum_2inact 0.67120
cond_oc_dum_3unemp 0.36200
susprindum_oh 0.82687
susprindum_coc 0.48131
susprindum_pbc 0.61062
susprindum_mar 0.20693
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.228 0.814 1.104 1.366
less_90d_tr1.lag 2.232 0.448 1.920 2.594
log_dias_treat_imp_sin_na.lag 1.193 0.838 1.110 1.282
comp_bpsc_y3_severe.lag 1.191 0.840 1.036 1.368
policonsumo2.lag 0.955 1.047 0.854 1.069
edad_al_ing_1 1.119 0.893 1.098 1.141
ano_nac_corr 1.126 0.888 1.105 1.147
susinidum_oh 1.038 0.963 0.787 1.371
susinidum_coc 1.091 0.917 0.795 1.496
susinidum_pbc 1.021 0.979 0.756 1.379
susinidum_mar 1.087 0.920 0.819 1.443
psycom_dum_with 0.975 1.025 0.875 1.087
psycom_dum_study 0.670 1.492 0.581 0.774
freq_cons_dum_5day 0.957 1.044 0.867 1.058
cond_oc_dum_2inact 1.039 0.963 0.871 1.239
cond_oc_dum_3unemp 1.086 0.920 0.909 1.298
susprindum_oh 0.970 1.031 0.735 1.279
susprindum_coc 0.904 1.107 0.682 1.198
susprindum_pbc 0.933 1.072 0.715 1.218
susprindum_mar 0.774 1.291 0.520 1.152
Concordance= 0.626 (se = 0.005 )
Likelihood ratio test= 328 on 20 df, p=<2e-16
Wald test = 374 on 20 df, p=<2e-16
Score (logrank) test = 315 on 20 df, p=<2e-16, Robust = 350 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.15 0.36 0.49 0.58 0.70 12.82
[1] "Modality: GP intensive ambulatory\nTime interval: [0,88]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 9973, number of events= 9973
(6233 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.20542 1.22804 0.03520 0.03199 6.42
less_90d_tr1.lag 1.03973 2.82845 0.03442 0.03813 27.26
log_dias_treat_imp_sin_na.lag 0.22867 1.25693 0.02066 0.02551 8.97
comp_bpsc_y3_severe.lag 0.36576 1.44161 0.02292 0.02274 16.09
policonsumo2.lag 0.16091 1.17458 0.03411 0.03307 4.87
edad_al_ing_1 0.13636 1.14609 0.00500 0.00450 30.29
ano_nac_corr 0.13706 1.14690 0.00492 0.00442 30.99
susinidum_oh 0.09086 1.09511 0.06911 0.06255 1.45
susinidum_coc 0.07104 1.07362 0.08273 0.08140 0.87
susinidum_pbc 0.17770 1.19447 0.07839 0.07475 2.38
susinidum_mar 0.17350 1.18946 0.07042 0.06400 2.71
psycom_dum_with -0.03422 0.96636 0.02230 0.02177 -1.57
psycom_dum_study -0.42746 0.65217 0.03271 0.03794 -11.27
freq_cons_dum_5day -0.00842 0.99161 0.02065 0.02115 -0.40
cond_oc_dum_2inact 0.00653 1.00655 0.03106 0.02998 0.22
cond_oc_dum_3unemp 0.01222 1.01230 0.02224 0.02276 0.54
susprindum_oh -0.06775 0.93449 0.08948 0.09220 -0.73
susprindum_coc -0.02664 0.97371 0.09056 0.09304 -0.29
susprindum_pbc -0.01332 0.98677 0.08833 0.09113 -0.15
susprindum_mar 0.01243 1.01251 0.09711 0.09863 0.13
Pr(>|z|)
tr_outcome.lag 1.3e-10 ***
less_90d_tr1.lag < 2e-16 ***
log_dias_treat_imp_sin_na.lag < 2e-16 ***
comp_bpsc_y3_severe.lag < 2e-16 ***
policonsumo2.lag 1.1e-06 ***
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 0.1464
susinidum_coc 0.3828
susinidum_pbc 0.0174 *
susinidum_mar 0.0067 **
psycom_dum_with 0.1160
psycom_dum_study < 2e-16 ***
freq_cons_dum_5day 0.6904
cond_oc_dum_2inact 0.8276
cond_oc_dum_3unemp 0.5913
susprindum_oh 0.4624
susprindum_coc 0.7746
susprindum_pbc 0.8838
susprindum_mar 0.8997
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.228 0.814 1.153 1.308
less_90d_tr1.lag 2.828 0.354 2.625 3.048
log_dias_treat_imp_sin_na.lag 1.257 0.796 1.196 1.321
comp_bpsc_y3_severe.lag 1.442 0.694 1.379 1.507
policonsumo2.lag 1.175 0.851 1.101 1.253
edad_al_ing_1 1.146 0.873 1.136 1.156
ano_nac_corr 1.147 0.872 1.137 1.157
susinidum_oh 1.095 0.913 0.969 1.238
susinidum_coc 1.074 0.931 0.915 1.259
susinidum_pbc 1.194 0.837 1.032 1.383
susinidum_mar 1.189 0.841 1.049 1.348
psycom_dum_with 0.966 1.035 0.926 1.008
psycom_dum_study 0.652 1.533 0.605 0.703
freq_cons_dum_5day 0.992 1.008 0.951 1.034
cond_oc_dum_2inact 1.007 0.993 0.949 1.067
cond_oc_dum_3unemp 1.012 0.988 0.968 1.058
susprindum_oh 0.934 1.070 0.780 1.120
susprindum_coc 0.974 1.027 0.811 1.169
susprindum_pbc 0.987 1.013 0.825 1.180
susprindum_mar 1.013 0.988 0.835 1.228
Concordance= 0.661 (se = 0.002 )
Likelihood ratio test= 2547 on 20 df, p=<2e-16
Wald test = 2470 on 20 df, p=<2e-16
Score (logrank) test = 2507 on 20 df, p=<2e-16, Robust = 2298 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.07 0.18 0.29 0.38 0.50 16.37
[1] "Modality: basic ambulatory\nTime interval: [0,88]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 8836, number of events= 8836
(5238 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.19144 1.21100 0.03883 0.03485 5.49
less_90d_tr1.lag 0.85480 2.35091 0.03764 0.04065 21.03
log_dias_treat_imp_sin_na.lag 0.15761 1.17071 0.02305 0.02568 6.14
comp_bpsc_y3_severe.lag 0.71607 2.04638 0.02649 0.02951 24.26
policonsumo2.lag 0.05970 1.06152 0.03302 0.02996 1.99
edad_al_ing_1 0.12723 1.13568 0.00543 0.00497 25.62
ano_nac_corr 0.12930 1.13803 0.00536 0.00489 26.45
susinidum_oh 0.06745 1.06978 0.07583 0.06475 1.04
susinidum_coc 0.10834 1.11443 0.08975 0.08886 1.22
susinidum_pbc 0.15305 1.16538 0.08700 0.07998 1.91
susinidum_mar 0.12643 1.13477 0.07701 0.06677 1.89
psycom_dum_with -0.09434 0.90997 0.02377 0.02360 -4.00
psycom_dum_study -0.22528 0.79829 0.03267 0.03752 -6.00
freq_cons_dum_5day -0.05778 0.94386 0.02458 0.02544 -2.27
cond_oc_dum_2inact -0.03710 0.96358 0.03369 0.03624 -1.02
cond_oc_dum_3unemp -0.01370 0.98639 0.02594 0.02723 -0.50
susprindum_oh 0.03720 1.03790 0.10027 0.07352 0.51
susprindum_coc 0.04959 1.05084 0.10158 0.07538 0.66
susprindum_pbc 0.04364 1.04461 0.10005 0.07375 0.59
susprindum_mar 0.08895 1.09303 0.10631 0.08072 1.10
Pr(>|z|)
tr_outcome.lag 3.9e-08 ***
less_90d_tr1.lag < 2e-16 ***
log_dias_treat_imp_sin_na.lag 8.3e-10 ***
comp_bpsc_y3_severe.lag < 2e-16 ***
policonsumo2.lag 0.046 *
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 0.298
susinidum_coc 0.223
susinidum_pbc 0.056 .
susinidum_mar 0.058 .
psycom_dum_with 6.4e-05 ***
psycom_dum_study 1.9e-09 ***
freq_cons_dum_5day 0.023 *
cond_oc_dum_2inact 0.306
cond_oc_dum_3unemp 0.615
susprindum_oh 0.613
susprindum_coc 0.511
susprindum_pbc 0.554
susprindum_mar 0.270
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.211 0.826 1.131 1.297
less_90d_tr1.lag 2.351 0.425 2.171 2.546
log_dias_treat_imp_sin_na.lag 1.171 0.854 1.113 1.231
comp_bpsc_y3_severe.lag 2.046 0.489 1.931 2.168
policonsumo2.lag 1.062 0.942 1.001 1.126
edad_al_ing_1 1.136 0.881 1.125 1.147
ano_nac_corr 1.138 0.879 1.127 1.149
susinidum_oh 1.070 0.935 0.942 1.215
susinidum_coc 1.114 0.897 0.936 1.326
susinidum_pbc 1.165 0.858 0.996 1.363
susinidum_mar 1.135 0.881 0.996 1.293
psycom_dum_with 0.910 1.099 0.869 0.953
psycom_dum_study 0.798 1.253 0.742 0.859
freq_cons_dum_5day 0.944 1.059 0.898 0.992
cond_oc_dum_2inact 0.964 1.038 0.898 1.035
cond_oc_dum_3unemp 0.986 1.014 0.935 1.040
susprindum_oh 1.038 0.963 0.899 1.199
susprindum_coc 1.051 0.952 0.907 1.218
susprindum_pbc 1.045 0.957 0.904 1.207
susprindum_mar 1.093 0.915 0.933 1.280
Concordance= 0.678 (se = 0.003 )
Likelihood ratio test= 2793 on 20 df, p=<2e-16
Wald test = 2423 on 20 df, p=<2e-16
Score (logrank) test = 2875 on 20 df, p=<2e-16, Robust = 2414 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.06 0.17 0.29 0.42 0.62 9.17
[1] "Modality: GP residential\nTime interval: [0,88]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 4436, number of events= 4436
(2815 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.30950 1.36275 0.04465 0.03996 7.75
less_90d_tr1.lag 0.77198 2.16405 0.04672 0.05172 14.93
log_dias_treat_imp_sin_na.lag 0.19686 1.21757 0.02213 0.02760 7.13
comp_bpsc_y3_severe.lag 0.33683 1.40050 0.04277 0.03970 8.49
policonsumo2.lag 0.21234 1.23657 0.05514 0.05263 4.03
edad_al_ing_1 0.10459 1.11026 0.00769 0.00679 15.41
ano_nac_corr 0.10628 1.11214 0.00765 0.00667 15.94
susinidum_oh 0.09816 1.10314 0.10611 0.10080 0.97
susinidum_coc 0.16421 1.17846 0.12980 0.12460 1.32
susinidum_pbc 0.23338 1.26286 0.11301 0.11103 2.10
susinidum_mar 0.24446 1.27694 0.10643 0.10096 2.42
psycom_dum_with -0.04315 0.95777 0.03427 0.03465 -1.25
psycom_dum_study -0.31271 0.73146 0.04962 0.05491 -5.69
freq_cons_dum_5day 0.01423 1.01433 0.03482 0.03425 0.42
cond_oc_dum_2inact 0.07737 1.08044 0.05583 0.05164 1.50
cond_oc_dum_3unemp 0.09295 1.09740 0.04599 0.04394 2.12
susprindum_oh 0.08714 1.09105 0.16038 0.15169 0.57
susprindum_coc -0.00799 0.99204 0.16159 0.15287 -0.05
susprindum_pbc -0.06583 0.93629 0.15656 0.14868 -0.44
susprindum_mar 0.10378 1.10935 0.18041 0.16729 0.62
Pr(>|z|)
tr_outcome.lag 9.5e-15 ***
less_90d_tr1.lag < 2e-16 ***
log_dias_treat_imp_sin_na.lag 9.9e-13 ***
comp_bpsc_y3_severe.lag < 2e-16 ***
policonsumo2.lag 5.5e-05 ***
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 0.330
susinidum_coc 0.188
susinidum_pbc 0.036 *
susinidum_mar 0.015 *
psycom_dum_with 0.213
psycom_dum_study 1.2e-08 ***
freq_cons_dum_5day 0.678
cond_oc_dum_2inact 0.134
cond_oc_dum_3unemp 0.034 *
susprindum_oh 0.566
susprindum_coc 0.958
susprindum_pbc 0.658
susprindum_mar 0.535
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.363 0.734 1.260 1.474
less_90d_tr1.lag 2.164 0.462 1.955 2.395
log_dias_treat_imp_sin_na.lag 1.218 0.821 1.153 1.285
comp_bpsc_y3_severe.lag 1.401 0.714 1.296 1.514
policonsumo2.lag 1.237 0.809 1.115 1.371
edad_al_ing_1 1.110 0.901 1.096 1.125
ano_nac_corr 1.112 0.899 1.098 1.127
susinidum_oh 1.103 0.907 0.905 1.344
susinidum_coc 1.178 0.849 0.923 1.504
susinidum_pbc 1.263 0.792 1.016 1.570
susinidum_mar 1.277 0.783 1.048 1.556
psycom_dum_with 0.958 1.044 0.895 1.025
psycom_dum_study 0.731 1.367 0.657 0.815
freq_cons_dum_5day 1.014 0.986 0.948 1.085
cond_oc_dum_2inact 1.080 0.926 0.976 1.196
cond_oc_dum_3unemp 1.097 0.911 1.007 1.196
susprindum_oh 1.091 0.917 0.810 1.469
susprindum_coc 0.992 1.008 0.735 1.339
susprindum_pbc 0.936 1.068 0.700 1.253
susprindum_mar 1.109 0.901 0.799 1.540
Concordance= 0.642 (se = 0.004 )
Likelihood ratio test= 828 on 20 df, p=<2e-16
Wald test = 790 on 20 df, p=<2e-16
Score (logrank) test = 780 on 20 df, p=<2e-16, Robust = 724 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.08 0.18 0.24 0.30 0.37 6.84
[1] "Modality: WO intensive ambulatory\nTime interval: [0,88]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1546, number of events= 1546
(908 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.23134 1.26029 0.09190 0.07725 2.99
less_90d_tr1.lag 0.95955 2.61051 0.08825 0.08096 11.85
log_dias_treat_imp_sin_na.lag 0.09043 1.09465 0.05335 0.04257 2.12
comp_bpsc_y3_severe.lag 0.31352 1.36824 0.06024 0.05549 5.65
policonsumo2.lag -0.08198 0.92129 0.08184 0.06548 -1.25
edad_al_ing_1 0.12309 1.13099 0.01311 0.01174 10.49
ano_nac_corr 0.12241 1.13022 0.01264 0.01153 10.61
susinidum_oh -0.19596 0.82204 0.17904 0.17621 -1.11
susinidum_coc 0.00617 1.00619 0.20987 0.20957 0.03
susinidum_pbc -0.06247 0.93945 0.19053 0.18657 -0.33
susinidum_mar -0.06877 0.93355 0.18001 0.17810 -0.39
psycom_dum_with -0.05898 0.94272 0.05788 0.05379 -1.10
psycom_dum_study -0.47231 0.62356 0.08757 0.09703 -4.87
freq_cons_dum_5day -0.03117 0.96931 0.05294 0.05513 -0.57
cond_oc_dum_2inact 0.11950 1.12693 0.06683 0.06368 1.88
cond_oc_dum_3unemp 0.02634 1.02669 0.06935 0.07437 0.35
susprindum_oh -0.09921 0.90555 0.19340 0.19563 -0.51
susprindum_coc 0.01020 1.01025 0.19593 0.21054 0.05
susprindum_pbc -0.07461 0.92811 0.18970 0.20223 -0.37
susprindum_mar -0.08083 0.92235 0.20716 0.21899 -0.37
Pr(>|z|)
tr_outcome.lag 0.0027 **
less_90d_tr1.lag < 2e-16 ***
log_dias_treat_imp_sin_na.lag 0.0336 *
comp_bpsc_y3_severe.lag 1.6e-08 ***
policonsumo2.lag 0.2106
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 0.2661
susinidum_coc 0.9765
susinidum_pbc 0.7378
susinidum_mar 0.6994
psycom_dum_with 0.2729
psycom_dum_study 1.1e-06 ***
freq_cons_dum_5day 0.5718
cond_oc_dum_2inact 0.0606 .
cond_oc_dum_3unemp 0.7232
susprindum_oh 0.6121
susprindum_coc 0.9614
susprindum_pbc 0.7122
susprindum_mar 0.7121
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.260 0.793 1.083 1.466
less_90d_tr1.lag 2.611 0.383 2.227 3.059
log_dias_treat_imp_sin_na.lag 1.095 0.914 1.007 1.190
comp_bpsc_y3_severe.lag 1.368 0.731 1.227 1.525
policonsumo2.lag 0.921 1.085 0.810 1.047
edad_al_ing_1 1.131 0.884 1.105 1.157
ano_nac_corr 1.130 0.885 1.105 1.156
susinidum_oh 0.822 1.216 0.582 1.161
susinidum_coc 1.006 0.994 0.667 1.517
susinidum_pbc 0.939 1.064 0.652 1.354
susinidum_mar 0.934 1.071 0.658 1.324
psycom_dum_with 0.943 1.061 0.848 1.048
psycom_dum_study 0.624 1.604 0.516 0.754
freq_cons_dum_5day 0.969 1.032 0.870 1.080
cond_oc_dum_2inact 1.127 0.887 0.995 1.277
cond_oc_dum_3unemp 1.027 0.974 0.887 1.188
susprindum_oh 0.906 1.104 0.617 1.329
susprindum_coc 1.010 0.990 0.669 1.526
susprindum_pbc 0.928 1.077 0.624 1.380
susprindum_mar 0.922 1.084 0.600 1.417
Concordance= 0.659 (se = 0.006 )
Likelihood ratio test= 396 on 20 df, p=<2e-16
Wald test = 391 on 20 df, p=<2e-16
Score (logrank) test = 404 on 20 df, p=<2e-16, Robust = 375 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.134 0.312 0.489 0.630 0.868 2.768
[1] "Modality: WO residential\nTime interval: (93,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 34, number of events= 34
(32 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag NA NA 0.0000 0.0000 NA
less_90d_tr1.lag 4.4157 82.7386 2.7593 2.7900 1.58
log_dias_treat_imp_sin_na.lag 4.4195 83.0584 2.1191 1.9017 2.32
comp_bpsc_y3_severe.lag NA NA 0.0000 0.0000 NA
policonsumo2.lag NA NA 0.0000 0.0000 NA
edad_al_ing_1 2.4852 12.0039 0.7005 0.8291 3.00
ano_nac_corr 2.3745 10.7460 0.6782 0.8038 2.95
susinidum_oh -4.3298 0.0132 1.6380 1.3810 -3.14
susinidum_coc NA NA 0.0000 0.0000 NA
susinidum_pbc -4.5556 0.0105 1.6927 1.5403 -2.96
susinidum_mar -4.0816 0.0169 1.5315 1.0705 -3.81
psycom_dum_with 0.3149 1.3702 0.6251 0.5934 0.53
psycom_dum_study 1.8290 6.2277 1.2291 1.1187 1.63
freq_cons_dum_5day -0.2795 0.7561 0.5600 0.6164 -0.45
cond_oc_dum_2inact 0.6222 1.8630 1.2392 1.7717 0.35
cond_oc_dum_3unemp 1.1566 3.1792 1.1545 1.6923 0.68
susprindum_oh -1.6687 0.1885 1.9275 1.7284 -0.97
susprindum_coc -0.5048 0.6036 1.4589 1.0889 -0.46
susprindum_pbc -0.5150 0.5975 1.2052 0.7966 -0.65
susprindum_mar NA NA 0.0000 0.0000 NA
Pr(>|z|)
tr_outcome.lag NA
less_90d_tr1.lag 0.11349
log_dias_treat_imp_sin_na.lag 0.02013 *
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 0.00272 **
ano_nac_corr 0.00313 **
susinidum_oh 0.00172 **
susinidum_coc NA
susinidum_pbc 0.00310 **
susinidum_mar 0.00014 ***
psycom_dum_with 0.59557
psycom_dum_study 0.10206
freq_cons_dum_5day 0.65023
cond_oc_dum_2inact 0.72545
cond_oc_dum_3unemp 0.49430
susprindum_oh 0.33431
susprindum_coc 0.64295
susprindum_pbc 0.51798
susprindum_mar NA
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag NA NA NA NA
less_90d_tr1.lag 82.7386 0.0121 0.349036 1.96e+04
log_dias_treat_imp_sin_na.lag 83.0584 0.0120 1.998148 3.45e+03
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 12.0039 0.0833 2.363494 6.10e+01
ano_nac_corr 10.7460 0.0931 2.223632 5.19e+01
susinidum_oh 0.0132 75.9321 0.000879 1.97e-01
susinidum_coc NA NA NA NA
susinidum_pbc 0.0105 95.1637 0.000513 2.15e-01
susinidum_mar 0.0169 59.2412 0.002071 1.38e-01
psycom_dum_with 1.3702 0.7298 0.428266 4.38e+00
psycom_dum_study 6.2277 0.1606 0.695181 5.58e+01
freq_cons_dum_5day 0.7561 1.3225 0.225893 2.53e+00
cond_oc_dum_2inact 1.8630 0.5368 0.057826 6.00e+01
cond_oc_dum_3unemp 3.1792 0.3145 0.115310 8.77e+01
susprindum_oh 0.1885 5.3052 0.006370 5.58e+00
susprindum_coc 0.6036 1.6566 0.071431 5.10e+00
susprindum_pbc 0.5975 1.6736 0.125403 2.85e+00
susprindum_mar NA NA NA NA
Concordance= 0.777 (se = 0.035 )
Likelihood ratio test= 29.1 on 15 df, p=0.02
Wald test = 93.5 on 15 df, p=2e-13
Score (logrank) test = 36.1 on 15 df, p=0.002, Robust = 28.5 p=0.02
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.02 0.19 0.63 1.20 1.94 4.91
[1] "Modality: GP intensive ambulatory\nTime interval: (93,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 213, number of events= 213
(202 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.1420 0.8676 0.9397 0.6795 -0.21
less_90d_tr1.lag 0.1767 1.1933 1.2881 0.9068 0.19
log_dias_treat_imp_sin_na.lag -0.6701 0.5116 0.9782 0.9147 -0.73
comp_bpsc_y3_severe.lag -0.0543 0.9471 0.6691 0.5533 -0.10
policonsumo2.lag -0.4395 0.6444 0.8116 0.5030 -0.87
edad_al_ing_1 0.7008 2.0153 0.1090 0.1087 6.45
ano_nac_corr 0.6961 2.0059 0.1104 0.1105 6.30
susinidum_oh -0.0531 0.9483 0.3644 0.3947 -0.13
susinidum_coc 0.0278 1.0282 0.5261 0.4879 0.06
susinidum_pbc 0.0657 1.0679 0.5213 0.6393 0.10
susinidum_mar -0.0596 0.9421 0.3790 0.4164 -0.14
psycom_dum_with -0.1269 0.8808 0.1630 0.1558 -0.81
psycom_dum_study -0.1664 0.8467 0.2567 0.2651 -0.63
freq_cons_dum_5day 0.2489 1.2826 0.1656 0.1562 1.59
cond_oc_dum_2inact -0.0389 0.9619 0.2675 0.2474 -0.16
cond_oc_dum_3unemp -0.0623 0.9396 0.1663 0.1661 -0.38
susprindum_oh -0.8567 0.4246 0.6738 0.5892 -1.45
susprindum_coc -0.9290 0.3950 0.6812 0.6043 -1.54
susprindum_pbc -0.5742 0.5631 0.6592 0.5839 -0.98
susprindum_mar -0.7788 0.4590 0.7109 0.6565 -1.19
Pr(>|z|)
tr_outcome.lag 0.83
less_90d_tr1.lag 0.85
log_dias_treat_imp_sin_na.lag 0.46
comp_bpsc_y3_severe.lag 0.92
policonsumo2.lag 0.38
edad_al_ing_1 1.1e-10 ***
ano_nac_corr 3.0e-10 ***
susinidum_oh 0.89
susinidum_coc 0.95
susinidum_pbc 0.92
susinidum_mar 0.89
psycom_dum_with 0.42
psycom_dum_study 0.53
freq_cons_dum_5day 0.11
cond_oc_dum_2inact 0.88
cond_oc_dum_3unemp 0.71
susprindum_oh 0.15
susprindum_coc 0.12
susprindum_pbc 0.33
susprindum_mar 0.24
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.868 1.153 0.2290 3.29
less_90d_tr1.lag 1.193 0.838 0.2018 7.06
log_dias_treat_imp_sin_na.lag 0.512 1.955 0.0852 3.07
comp_bpsc_y3_severe.lag 0.947 1.056 0.3202 2.80
policonsumo2.lag 0.644 1.552 0.2404 1.73
edad_al_ing_1 2.015 0.496 1.6287 2.49
ano_nac_corr 2.006 0.499 1.6152 2.49
susinidum_oh 0.948 1.054 0.4375 2.06
susinidum_coc 1.028 0.973 0.3951 2.68
susinidum_pbc 1.068 0.936 0.3051 3.74
susinidum_mar 0.942 1.061 0.4165 2.13
psycom_dum_with 0.881 1.135 0.6490 1.20
psycom_dum_study 0.847 1.181 0.5036 1.42
freq_cons_dum_5day 1.283 0.780 0.9443 1.74
cond_oc_dum_2inact 0.962 1.040 0.5923 1.56
cond_oc_dum_3unemp 0.940 1.064 0.6785 1.30
susprindum_oh 0.425 2.355 0.1338 1.35
susprindum_coc 0.395 2.532 0.1208 1.29
susprindum_pbc 0.563 1.776 0.1793 1.77
susprindum_mar 0.459 2.179 0.1268 1.66
Concordance= 0.648 (se = 0.021 )
Likelihood ratio test= 62.5 on 20 df, p=3e-06
Wald test = 74.1 on 20 df, p=4e-08
Score (logrank) test = 58 on 20 df, p=1e-05, Robust = 60.1 p=7e-06
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.57 2.31 3.29 4.18 4.76 29.49
[1] "Modality: basic ambulatory\nTime interval: (93,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 200, number of events= 200
(189 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.7197 0.4869 1.2028 0.3990 -1.80
less_90d_tr1.lag -1.8129 0.1632 1.7882 1.2190 -1.49
log_dias_treat_imp_sin_na.lag -1.8079 0.1640 1.0549 0.8626 -2.10
comp_bpsc_y3_severe.lag 0.0974 1.1023 1.1447 0.4256 0.23
policonsumo2.lag 2.2023 9.0461 1.9113 1.0858 2.03
edad_al_ing_1 0.7794 2.1801 0.1172 0.1249 6.24
ano_nac_corr 0.7680 2.1554 0.1169 0.1245 6.17
susinidum_oh 0.5302 1.6993 0.5270 0.3807 1.39
susinidum_coc 0.1994 1.2206 0.5865 0.4163 0.48
susinidum_pbc 0.6580 1.9309 0.5718 0.4256 1.55
susinidum_mar 0.7490 2.1148 0.5406 0.3995 1.87
psycom_dum_with 0.0195 1.0197 0.1761 0.1662 0.12
psycom_dum_study 0.0389 1.0397 0.2107 0.2057 0.19
freq_cons_dum_5day 0.2773 1.3195 0.1653 0.1702 1.63
cond_oc_dum_2inact 0.2667 1.3057 0.2393 0.2809 0.95
cond_oc_dum_3unemp -0.0170 0.9831 0.1848 0.1678 -0.10
susprindum_oh -0.8348 0.4340 0.5167 0.4450 -1.88
susprindum_coc -0.8892 0.4110 0.5240 0.4586 -1.94
susprindum_pbc -0.7822 0.4574 0.5047 0.4376 -1.79
susprindum_mar -0.3340 0.7161 0.5473 0.5222 -0.64
Pr(>|z|)
tr_outcome.lag 0.071 .
less_90d_tr1.lag 0.137
log_dias_treat_imp_sin_na.lag 0.036 *
comp_bpsc_y3_severe.lag 0.819
policonsumo2.lag 0.043 *
edad_al_ing_1 4.3e-10 ***
ano_nac_corr 7.0e-10 ***
susinidum_oh 0.164
susinidum_coc 0.632
susinidum_pbc 0.122
susinidum_mar 0.061 .
psycom_dum_with 0.907
psycom_dum_study 0.850
freq_cons_dum_5day 0.103
cond_oc_dum_2inact 0.342
cond_oc_dum_3unemp 0.919
susprindum_oh 0.061 .
susprindum_coc 0.052 .
susprindum_pbc 0.074 .
susprindum_mar 0.522
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.487 2.054 0.2228 1.064
less_90d_tr1.lag 0.163 6.128 0.0150 1.779
log_dias_treat_imp_sin_na.lag 0.164 6.098 0.0302 0.889
comp_bpsc_y3_severe.lag 1.102 0.907 0.4787 2.538
policonsumo2.lag 9.046 0.111 1.0769 75.986
edad_al_ing_1 2.180 0.459 1.7068 2.785
ano_nac_corr 2.155 0.464 1.6885 2.751
susinidum_oh 1.699 0.588 0.8057 3.584
susinidum_coc 1.221 0.819 0.5398 2.760
susinidum_pbc 1.931 0.518 0.8384 4.447
susinidum_mar 2.115 0.473 0.9666 4.627
psycom_dum_with 1.020 0.981 0.7361 1.412
psycom_dum_study 1.040 0.962 0.6948 1.556
freq_cons_dum_5day 1.320 0.758 0.9453 1.842
cond_oc_dum_2inact 1.306 0.766 0.7529 2.264
cond_oc_dum_3unemp 0.983 1.017 0.7075 1.366
susprindum_oh 0.434 2.304 0.1814 1.038
susprindum_coc 0.411 2.433 0.1673 1.010
susprindum_pbc 0.457 2.186 0.1940 1.078
susprindum_mar 0.716 1.397 0.2573 1.993
Concordance= 0.664 (se = 0.022 )
Likelihood ratio test= 68.2 on 20 df, p=4e-07
Wald test = 175 on 20 df, p=<2e-16
Score (logrank) test = 66.3 on 20 df, p=7e-07, Robust = 66.8 p=6e-07
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.24 0.70 1.13 1.58 1.96 11.84
[1] "Modality: GP residential\nTime interval: (93,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 87, number of events= 87
(80 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -2.5530 0.0779 2.0164 0.9029 -2.83
less_90d_tr1.lag 2.0565 7.8187 1.0277 0.5534 3.72
log_dias_treat_imp_sin_na.lag 0.5230 1.6871 1.4091 0.6244 0.84
comp_bpsc_y3_severe.lag -0.0445 0.9565 1.7827 0.8678 -0.05
policonsumo2.lag -2.8354 0.0587 2.0626 1.3710 -2.07
edad_al_ing_1 1.0157 2.7612 0.2224 0.1885 5.39
ano_nac_corr 1.0052 2.7325 0.2221 0.1868 5.38
susinidum_oh -0.9494 0.3870 1.0531 0.3347 -2.84
susinidum_coc -0.9051 0.4045 1.2157 0.5140 -1.76
susinidum_pbc -1.3499 0.2593 1.1023 0.4586 -2.94
susinidum_mar -1.1042 0.3315 1.0590 0.3061 -3.61
psycom_dum_with -0.0621 0.9398 0.2946 0.2650 -0.23
psycom_dum_study 0.2976 1.3466 0.4129 0.3391 0.88
freq_cons_dum_5day -0.7801 0.4583 0.2954 0.2557 -3.05
cond_oc_dum_2inact -0.2235 0.7997 0.5043 0.5206 -0.43
cond_oc_dum_3unemp -0.5810 0.5594 0.4172 0.4796 -1.21
susprindum_oh -3.2829 0.0375 1.2809 0.6779 -4.84
susprindum_coc -2.3737 0.0931 1.2918 0.8134 -2.92
susprindum_pbc -2.1977 0.1111 1.2660 0.7348 -2.99
susprindum_mar -1.6220 0.1975 1.5119 0.7975 -2.03
Pr(>|z|)
tr_outcome.lag 0.00469 **
less_90d_tr1.lag 0.00020 ***
log_dias_treat_imp_sin_na.lag 0.40220
comp_bpsc_y3_severe.lag 0.95912
policonsumo2.lag 0.03863 *
edad_al_ing_1 7.1e-08 ***
ano_nac_corr 7.4e-08 ***
susinidum_oh 0.00457 **
susinidum_coc 0.07827 .
susinidum_pbc 0.00324 **
susinidum_mar 0.00031 ***
psycom_dum_with 0.81465
psycom_dum_study 0.38018
freq_cons_dum_5day 0.00228 **
cond_oc_dum_2inact 0.66772
cond_oc_dum_3unemp 0.22571
susprindum_oh 1.3e-06 ***
susprindum_coc 0.00352 **
susprindum_pbc 0.00278 **
susprindum_mar 0.04197 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.0779 12.845 0.01326 0.457
less_90d_tr1.lag 7.8187 0.128 2.64266 23.133
log_dias_treat_imp_sin_na.lag 1.6871 0.593 0.49623 5.736
comp_bpsc_y3_severe.lag 0.9565 1.045 0.17460 5.240
policonsumo2.lag 0.0587 17.037 0.00400 0.862
edad_al_ing_1 2.7612 0.362 1.90834 3.995
ano_nac_corr 2.7325 0.366 1.89481 3.941
susinidum_oh 0.3870 2.584 0.20079 0.746
susinidum_coc 0.4045 2.472 0.14770 1.108
susinidum_pbc 0.2593 3.857 0.10554 0.637
susinidum_mar 0.3315 3.017 0.18192 0.604
psycom_dum_with 0.9398 1.064 0.55910 1.580
psycom_dum_study 1.3466 0.743 0.69279 2.617
freq_cons_dum_5day 0.4583 2.182 0.27768 0.757
cond_oc_dum_2inact 0.7997 1.250 0.28831 2.218
cond_oc_dum_3unemp 0.5594 1.788 0.21852 1.432
susprindum_oh 0.0375 26.654 0.00994 0.142
susprindum_coc 0.0931 10.737 0.01891 0.459
susprindum_pbc 0.1111 9.004 0.02631 0.469
susprindum_mar 0.1975 5.063 0.04138 0.943
Concordance= 0.718 (se = 0.029 )
Likelihood ratio test= 47.9 on 20 df, p=4e-04
Wald test = 120 on 20 df, p=3e-16
Score (logrank) test = 47.1 on 20 df, p=6e-04, Robust = 42.4 p=0.002
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
334 1060 1807 3592 3996 25185
[1] "Modality: WO intensive ambulatory\nTime interval: (93,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 41, number of events= 41
(36 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 1.20e+01 1.65e+05 4.55e+03 8.85e-01 13.58
less_90d_tr1.lag 2.41e+01 3.00e+10 7.73e+03 2.50e+00 9.64
log_dias_treat_imp_sin_na.lag -2.91e+01 2.33e-13 1.15e+04 2.34e+00 -12.44
comp_bpsc_y3_severe.lag -3.46e+01 9.18e-16 1.26e+04 2.68e+00 -12.94
policonsumo2.lag -7.42e+01 5.80e-33 2.82e+04 5.70e+00 -13.02
edad_al_ing_1 3.98e-01 1.49e+00 3.04e-01 3.27e-01 1.22
ano_nac_corr 3.63e-01 1.44e+00 2.94e-01 3.20e-01 1.14
susinidum_oh -1.29e+00 2.76e-01 6.11e-01 7.17e-01 -1.79
susinidum_coc -3.32e+00 3.62e-02 1.39e+00 1.02e+00 -3.25
susinidum_pbc -1.46e+00 2.32e-01 7.40e-01 8.96e-01 -1.63
susinidum_mar NA NA 0.00e+00 0.00e+00 NA
psycom_dum_with 4.43e-01 1.56e+00 5.68e-01 5.14e-01 0.86
psycom_dum_study -3.73e-01 6.88e-01 6.21e-01 6.65e-01 -0.56
freq_cons_dum_5day 1.94e+00 6.93e+00 6.29e-01 6.19e-01 3.13
cond_oc_dum_2inact -1.71e+00 1.80e-01 9.74e-01 1.08e+00 -1.58
cond_oc_dum_3unemp -1.81e+00 1.63e-01 1.05e+00 1.06e+00 -1.72
susprindum_oh -1.79e+00 1.68e-01 1.55e+00 1.52e+00 -1.18
susprindum_coc -5.81e-01 5.59e-01 1.26e+00 7.71e-01 -0.75
susprindum_pbc -5.69e-01 5.66e-01 1.18e+00 7.55e-01 -0.75
susprindum_mar -1.73e+00 1.77e-01 1.50e+00 1.11e+00 -1.56
Pr(>|z|)
tr_outcome.lag <2e-16 ***
less_90d_tr1.lag <2e-16 ***
log_dias_treat_imp_sin_na.lag <2e-16 ***
comp_bpsc_y3_severe.lag <2e-16 ***
policonsumo2.lag <2e-16 ***
edad_al_ing_1 0.2238
ano_nac_corr 0.2561
susinidum_oh 0.0727 .
susinidum_coc 0.0012 **
susinidum_pbc 0.1026
susinidum_mar NA
psycom_dum_with 0.3892
psycom_dum_study 0.5745
freq_cons_dum_5day 0.0018 **
cond_oc_dum_2inact 0.1142
cond_oc_dum_3unemp 0.0862 .
susprindum_oh 0.2387
susprindum_coc 0.4508
susprindum_pbc 0.4510
susprindum_mar 0.1199
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.65e+05 6.06e-06 2.92e+04 9.35e+05
less_90d_tr1.lag 3.00e+10 3.34e-11 2.22e+08 4.04e+12
log_dias_treat_imp_sin_na.lag 2.33e-13 4.29e+12 2.38e-15 2.28e-11
comp_bpsc_y3_severe.lag 9.18e-16 1.09e+15 4.85e-18 1.74e-13
policonsumo2.lag 5.80e-33 1.72e+32 8.14e-38 4.13e-28
edad_al_ing_1 1.49e+00 6.72e-01 7.84e-01 2.83e+00
ano_nac_corr 1.44e+00 6.96e-01 7.68e-01 2.69e+00
susinidum_oh 2.76e-01 3.62e+00 6.78e-02 1.13e+00
susinidum_coc 3.62e-02 2.76e+01 4.90e-03 2.68e-01
susinidum_pbc 2.32e-01 4.31e+00 4.01e-02 1.34e+00
susinidum_mar NA NA NA NA
psycom_dum_with 1.56e+00 6.42e-01 5.68e-01 4.27e+00
psycom_dum_study 6.88e-01 1.45e+00 1.87e-01 2.53e+00
freq_cons_dum_5day 6.93e+00 1.44e-01 2.06e+00 2.33e+01
cond_oc_dum_2inact 1.80e-01 5.55e+00 2.15e-02 1.51e+00
cond_oc_dum_3unemp 1.63e-01 6.12e+00 2.06e-02 1.29e+00
susprindum_oh 1.68e-01 5.96e+00 8.59e-03 3.27e+00
susprindum_coc 5.59e-01 1.79e+00 1.24e-01 2.53e+00
susprindum_pbc 5.66e-01 1.77e+00 1.29e-01 2.49e+00
susprindum_mar 1.77e-01 5.64e+00 2.00e-02 1.57e+00
Concordance= 0.725 (se = 0.05 )
Likelihood ratio test= 24.1 on 19 df, p=0.2
Wald test = 941 on 19 df, p=<2e-16
Score (logrank) test = 26.8 on 19 df, p=0.1, Robust = 19.6 p=0.4
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.62e+29 1.94e+30 4.51e+30 2.39e+37 9.76e+30 9.82e+38
[1] "Modality: WO residential\nTime interval: (88,93]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 13, number of events= 13
(13 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag NA NA 0.00e+00 0.00e+00 NA
less_90d_tr1.lag NA NA 0.00e+00 0.00e+00 NA
log_dias_treat_imp_sin_na.lag NA NA 0.00e+00 0.00e+00 NA
comp_bpsc_y3_severe.lag NA NA 0.00e+00 0.00e+00 NA
policonsumo2.lag NA NA 0.00e+00 0.00e+00 NA
edad_al_ing_1 -5.69e+01 2.01e-25 5.83e+02 2.31e+00 -24.7
ano_nac_corr -3.70e+01 8.29e-17 2.42e+04 1.94e+00 -19.1
susinidum_oh 2.40e+02 1.77e+104 3.87e+05 7.01e+00 34.2
susinidum_coc NA NA 0.00e+00 0.00e+00 NA
susinidum_pbc 1.68e+02 9.89e+72 1.69e+05 5.26e+00 31.9
susinidum_mar NA NA 0.00e+00 0.00e+00 NA
psycom_dum_with -2.17e+01 3.76e-10 7.26e+04 1.38e+00 -15.7
psycom_dum_study -1.21e+02 1.81e-53 4.11e+05 5.48e+00 -22.2
freq_cons_dum_5day -6.09e+01 3.71e-27 2.42e+04 2.26e+00 -26.9
cond_oc_dum_2inact NA NA 0.00e+00 0.00e+00 NA
cond_oc_dum_3unemp NA NA 0.00e+00 0.00e+00 NA
susprindum_oh 9.63e+01 6.38e+41 1.21e+05 5.27e+00 18.2
susprindum_coc 1.21e+02 2.66e+52 7.26e+04 3.47e+00 34.8
susprindum_pbc 1.14e+02 3.44e+49 2.43e+04 4.27e+00 26.7
susprindum_mar NA NA 0.00e+00 0.00e+00 NA
Pr(>|z|)
tr_outcome.lag NA
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag NA
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 <2e-16 ***
ano_nac_corr <2e-16 ***
susinidum_oh <2e-16 ***
susinidum_coc NA
susinidum_pbc <2e-16 ***
susinidum_mar NA
psycom_dum_with <2e-16 ***
psycom_dum_study <2e-16 ***
freq_cons_dum_5day <2e-16 ***
cond_oc_dum_2inact NA
cond_oc_dum_3unemp NA
susprindum_oh <2e-16 ***
susprindum_coc <2e-16 ***
susprindum_pbc <2e-16 ***
susprindum_mar NA
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag NA NA NA NA
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag NA NA NA NA
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 2.01e-25 4.98e+24 2.19e-27 1.84e-23
ano_nac_corr 8.29e-17 1.21e+16 1.84e-18 3.73e-15
susinidum_oh 1.77e+104 5.64e-105 1.92e+98 1.64e+110
susinidum_coc NA NA NA NA
susinidum_pbc 9.89e+72 1.01e-73 3.30e+68 2.97e+77
susinidum_mar NA NA NA NA
psycom_dum_with 3.76e-10 2.66e+09 2.52e-11 5.61e-09
psycom_dum_study 1.81e-53 5.53e+52 3.91e-58 8.35e-49
freq_cons_dum_5day 3.71e-27 2.70e+26 4.43e-29 3.11e-25
cond_oc_dum_2inact NA NA NA NA
cond_oc_dum_3unemp NA NA NA NA
susprindum_oh 6.38e+41 1.57e-42 2.07e+37 1.97e+46
susprindum_coc 2.66e+52 3.76e-53 2.98e+49 2.37e+55
susprindum_pbc 3.44e+49 2.90e-50 7.99e+45 1.48e+53
susprindum_mar NA NA NA NA
Concordance= 1 (se = 0 )
Likelihood ratio test= 45.1 on 10 df, p=2e-06
Wald test = 1819 on 10 df, p=<2e-16
Score (logrank) test = 16.9 on 10 df, p=0.08, Robust = 12.6 p=0.2
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.00e+00 0.00e+00 0.00e+00 2.29e+68 0.00e+00 2.98e+69
[1] "Modality: GP intensive ambulatory\nTime interval: (88,93]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 74, number of events= 74
(74 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag NA NA 0.0000 0.0000 NA
less_90d_tr1.lag NA NA 0.0000 0.0000 NA
log_dias_treat_imp_sin_na.lag NA NA 0.0000 0.0000 NA
comp_bpsc_y3_severe.lag NA NA 0.0000 0.0000 NA
policonsumo2.lag NA NA 0.0000 0.0000 NA
edad_al_ing_1 0.0225 1.0228 0.1475 0.1616 0.14
ano_nac_corr 0.0460 1.0471 0.1395 0.1496 0.31
susinidum_oh -0.2291 0.7953 0.2985 0.2714 -0.84
susinidum_coc 0.8542 2.3495 0.5729 0.5836 1.46
susinidum_pbc -0.6890 0.5021 0.5586 0.6669 -1.03
susinidum_mar NA NA 0.0000 0.0000 NA
psycom_dum_with 0.3664 1.4425 0.3078 0.2991 1.22
psycom_dum_study 1.1222 3.0717 0.4851 0.4892 2.29
freq_cons_dum_5day 0.2846 1.3292 0.3179 0.2895 0.98
cond_oc_dum_2inact 0.1720 1.1877 0.4223 0.3660 0.47
cond_oc_dum_3unemp 0.4730 1.6048 0.3933 0.3670 1.29
susprindum_oh -0.0812 0.9220 0.6170 0.4583 -0.18
susprindum_coc -0.0655 0.9366 0.6616 0.5184 -0.13
susprindum_pbc -0.6944 0.4994 0.5289 0.3637 -1.91
susprindum_mar NA NA 0.0000 0.0000 NA
Pr(>|z|)
tr_outcome.lag NA
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag NA
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 0.889
ano_nac_corr 0.758
susinidum_oh 0.399
susinidum_coc 0.143
susinidum_pbc 0.302
susinidum_mar NA
psycom_dum_with 0.221
psycom_dum_study 0.022 *
freq_cons_dum_5day 0.326
cond_oc_dum_2inact 0.638
cond_oc_dum_3unemp 0.197
susprindum_oh 0.859
susprindum_coc 0.900
susprindum_pbc 0.056 .
susprindum_mar NA
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag NA NA NA NA
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag NA NA NA NA
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 1.023 0.978 0.745 1.40
ano_nac_corr 1.047 0.955 0.781 1.40
susinidum_oh 0.795 1.257 0.467 1.35
susinidum_coc 2.350 0.426 0.749 7.37
susinidum_pbc 0.502 1.992 0.136 1.86
susinidum_mar NA NA NA NA
psycom_dum_with 1.443 0.693 0.803 2.59
psycom_dum_study 3.072 0.326 1.177 8.01
freq_cons_dum_5day 1.329 0.752 0.754 2.34
cond_oc_dum_2inact 1.188 0.842 0.580 2.43
cond_oc_dum_3unemp 1.605 0.623 0.782 3.29
susprindum_oh 0.922 1.085 0.375 2.26
susprindum_coc 0.937 1.068 0.339 2.59
susprindum_pbc 0.499 2.002 0.245 1.02
susprindum_mar NA NA NA NA
Concordance= 0.634 (se = 0.034 )
Likelihood ratio test= 13.9 on 13 df, p=0.4
Wald test = 23.4 on 13 df, p=0.04
Score (logrank) test = 14.5 on 13 df, p=0.3, Robust = 16.6 p=0.2
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.198 0.615 0.804 0.936 1.277 2.215
[1] "Modality: basic ambulatory\nTime interval: (88,93]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 77, number of events= 77
(76 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 1.31e+00 3.70e+00 1.22e+00 5.87e-01 2.23
less_90d_tr1.lag NA NA 0.00e+00 0.00e+00 NA
log_dias_treat_imp_sin_na.lag NA NA 0.00e+00 0.00e+00 NA
comp_bpsc_y3_severe.lag NA NA 0.00e+00 0.00e+00 NA
policonsumo2.lag NA NA 0.00e+00 0.00e+00 NA
edad_al_ing_1 -8.12e-02 9.22e-01 1.57e-01 1.39e-01 -0.58
ano_nac_corr -1.72e-01 8.42e-01 1.59e-01 1.40e-01 -1.23
susinidum_oh -6.05e-01 5.46e-01 6.92e-01 4.73e-01 -1.28
susinidum_coc -8.26e-02 9.21e-01 1.05e+00 6.08e-01 -0.14
susinidum_pbc -1.81e+00 1.64e-01 8.99e-01 6.97e-01 -2.59
susinidum_mar -8.53e-01 4.26e-01 7.36e-01 5.27e-01 -1.62
psycom_dum_with -3.47e-01 7.07e-01 3.23e-01 2.73e-01 -1.27
psycom_dum_study -6.93e-01 5.00e-01 3.65e-01 3.41e-01 -2.03
freq_cons_dum_5day -2.08e-01 8.12e-01 3.02e-01 2.87e-01 -0.73
cond_oc_dum_2inact 7.39e-01 2.09e+00 4.50e-01 4.40e-01 1.68
cond_oc_dum_3unemp 4.68e-01 1.60e+00 3.39e-01 3.44e-01 1.36
susprindum_oh -2.10e+01 7.35e-10 3.08e+03 6.52e-01 -32.26
susprindum_coc -2.08e+01 9.01e-10 3.08e+03 5.09e-01 -40.94
susprindum_pbc -1.95e+01 3.43e-09 3.08e+03 5.67e-01 -34.39
susprindum_mar -2.06e+01 1.17e-09 3.08e+03 6.89e-01 -29.86
Pr(>|z|)
tr_outcome.lag 0.0258 *
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag NA
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 0.5597
ano_nac_corr 0.2186
susinidum_oh 0.2008
susinidum_coc 0.8919
susinidum_pbc 0.0095 **
susinidum_mar 0.1056
psycom_dum_with 0.2028
psycom_dum_study 0.0423 *
freq_cons_dum_5day 0.4680
cond_oc_dum_2inact 0.0926 .
cond_oc_dum_3unemp 0.1732
susprindum_oh <2e-16 ***
susprindum_coc <2e-16 ***
susprindum_pbc <2e-16 ***
susprindum_mar <2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 3.70e+00 2.70e-01 1.17e+00 1.17e+01
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag NA NA NA NA
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 9.22e-01 1.08e+00 7.02e-01 1.21e+00
ano_nac_corr 8.42e-01 1.19e+00 6.41e-01 1.11e+00
susinidum_oh 5.46e-01 1.83e+00 2.16e-01 1.38e+00
susinidum_coc 9.21e-01 1.09e+00 2.80e-01 3.03e+00
susinidum_pbc 1.64e-01 6.11e+00 4.17e-02 6.43e-01
susinidum_mar 4.26e-01 2.35e+00 1.52e-01 1.20e+00
psycom_dum_with 7.07e-01 1.41e+00 4.14e-01 1.21e+00
psycom_dum_study 5.00e-01 2.00e+00 2.56e-01 9.76e-01
freq_cons_dum_5day 8.12e-01 1.23e+00 4.63e-01 1.42e+00
cond_oc_dum_2inact 2.09e+00 4.77e-01 8.85e-01 4.96e+00
cond_oc_dum_3unemp 1.60e+00 6.26e-01 8.14e-01 3.13e+00
susprindum_oh 7.35e-10 1.36e+09 2.05e-10 2.64e-09
susprindum_coc 9.01e-10 1.11e+09 3.33e-10 2.44e-09
susprindum_pbc 3.43e-09 2.91e+08 1.13e-09 1.04e-08
susprindum_mar 1.17e-09 8.58e+08 3.02e-10 4.50e-09
Concordance= 0.731 (se = 0.03 )
Likelihood ratio test= 44.5 on 16 df, p=2e-04
Wald test = 16181 on 16 df, p=<2e-16
Score (logrank) test = 111 on 16 df, p=2e-16, Robust = 37.4 p=0.002
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.00e+00 1.97e+08 3.29e+08 5.05e+08 7.07e+08 1.97e+09
[1] "Modality: GP residential\nTime interval: (88,93]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 44, number of events= 44
(44 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag NA NA 0.0000 0.0000 NA
less_90d_tr1.lag NA NA 0.0000 0.0000 NA
log_dias_treat_imp_sin_na.lag NA NA 0.0000 0.0000 NA
comp_bpsc_y3_severe.lag NA NA 0.0000 0.0000 NA
policonsumo2.lag NA NA 0.0000 0.0000 NA
edad_al_ing_1 -0.0941 0.9102 0.2559 0.2579 -0.36
ano_nac_corr -0.1054 0.9000 0.2603 0.2672 -0.39
susinidum_oh -1.3805 0.2515 1.1947 0.7912 -1.74
susinidum_coc -2.3484 0.0955 1.7688 1.0683 -2.20
susinidum_pbc -1.7372 0.1760 1.2633 0.9136 -1.90
susinidum_mar -1.7400 0.1755 1.2540 0.9334 -1.86
psycom_dum_with -0.5855 0.5568 0.4569 0.4924 -1.19
psycom_dum_study -0.5570 0.5729 0.7513 0.6086 -0.92
freq_cons_dum_5day 0.2100 1.2337 0.4609 0.4900 0.43
cond_oc_dum_2inact 1.6085 4.9952 0.8868 0.6243 2.58
cond_oc_dum_3unemp 0.6958 2.0053 0.6945 0.7280 0.96
susprindum_oh -3.2052 0.0406 1.4883 1.2502 -2.56
susprindum_coc -2.1759 0.1135 1.4852 1.1353 -1.92
susprindum_pbc -2.7616 0.0632 1.2925 0.9600 -2.88
susprindum_mar NA NA 0.0000 0.0000 NA
Pr(>|z|)
tr_outcome.lag NA
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag NA
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 0.715
ano_nac_corr 0.693
susinidum_oh 0.081 .
susinidum_coc 0.028 *
susinidum_pbc 0.057 .
susinidum_mar 0.062 .
psycom_dum_with 0.234
psycom_dum_study 0.360
freq_cons_dum_5day 0.668
cond_oc_dum_2inact 0.010 **
cond_oc_dum_3unemp 0.339
susprindum_oh 0.010 *
susprindum_coc 0.055 .
susprindum_pbc 0.004 **
susprindum_mar NA
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag NA NA NA NA
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag NA NA NA NA
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 0.9102 1.099 0.54904 1.509
ano_nac_corr 0.9000 1.111 0.53305 1.520
susinidum_oh 0.2515 3.977 0.05333 1.186
susinidum_coc 0.0955 10.469 0.01177 0.775
susinidum_pbc 0.1760 5.682 0.02937 1.055
susinidum_mar 0.1755 5.697 0.02817 1.094
psycom_dum_with 0.5568 1.796 0.21212 1.462
psycom_dum_study 0.5729 1.745 0.17380 1.889
freq_cons_dum_5day 1.2337 0.811 0.47219 3.223
cond_oc_dum_2inact 4.9952 0.200 1.46948 16.980
cond_oc_dum_3unemp 2.0053 0.499 0.48140 8.353
susprindum_oh 0.0406 24.660 0.00350 0.470
susprindum_coc 0.1135 8.810 0.01227 1.050
susprindum_pbc 0.0632 15.825 0.00963 0.415
susprindum_mar NA NA NA NA
Concordance= 0.62 (se = 0.051 )
Likelihood ratio test= 12.3 on 14 df, p=0.6
Wald test = 38.5 on 14 df, p=4e-04
Score (logrank) test = 20 on 14 df, p=0.1, Robust = 11.2 p=0.7
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
2.6 34.2 48.9 52.1 70.7 99.9
[1] "Modality: WO intensive ambulatory\nTime interval: (88,93]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 10, number of events= 10
(10 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag NA NA 0.00e+00 0.00e+00 NA
less_90d_tr1.lag NA NA 0.00e+00 0.00e+00 NA
log_dias_treat_imp_sin_na.lag NA NA 0.00e+00 0.00e+00 NA
comp_bpsc_y3_severe.lag NA NA 0.00e+00 0.00e+00 NA
policonsumo2.lag NA NA 0.00e+00 0.00e+00 NA
edad_al_ing_1 -5.34e+01 6.69e-24 6.64e+03 1.45e+00 -36.92
ano_nac_corr -4.99e+01 2.05e-22 5.79e+03 1.40e+00 -35.58
susinidum_oh 5.70e+00 2.99e+02 1.73e+03 1.07e+00 5.31
susinidum_coc NA NA 0.00e+00 0.00e+00 NA
susinidum_pbc 5.46e+01 5.41e+23 1.65e+04 3.47e+00 15.75
susinidum_mar NA NA 0.00e+00 0.00e+00 NA
psycom_dum_with 2.74e+01 8.14e+11 4.34e+03 1.45e+00 18.95
psycom_dum_study NA NA 0.00e+00 0.00e+00 NA
freq_cons_dum_5day -1.48e+02 6.27e-65 1.42e+04 3.58e+00 -41.25
cond_oc_dum_2inact -4.10e+01 1.49e-18 8.48e+03 2.02e+00 -20.30
cond_oc_dum_3unemp NA NA 0.00e+00 0.00e+00 NA
susprindum_oh 7.39e+01 1.23e+32 6.98e+03 3.27e+00 22.57
susprindum_coc 1.39e+02 1.89e+60 6.56e+03 3.67e+00 37.84
susprindum_pbc NA NA 0.00e+00 0.00e+00 NA
susprindum_mar NA NA 0.00e+00 0.00e+00 NA
Pr(>|z|)
tr_outcome.lag NA
less_90d_tr1.lag NA
log_dias_treat_imp_sin_na.lag NA
comp_bpsc_y3_severe.lag NA
policonsumo2.lag NA
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 1.1e-07 ***
susinidum_coc NA
susinidum_pbc < 2e-16 ***
susinidum_mar NA
psycom_dum_with < 2e-16 ***
psycom_dum_study NA
freq_cons_dum_5day < 2e-16 ***
cond_oc_dum_2inact < 2e-16 ***
cond_oc_dum_3unemp NA
susprindum_oh < 2e-16 ***
susprindum_coc < 2e-16 ***
susprindum_pbc NA
susprindum_mar NA
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag NA NA NA NA
less_90d_tr1.lag NA NA NA NA
log_dias_treat_imp_sin_na.lag NA NA NA NA
comp_bpsc_y3_severe.lag NA NA NA NA
policonsumo2.lag NA NA NA NA
edad_al_ing_1 6.69e-24 1.50e+23 3.94e-25 1.14e-22
ano_nac_corr 2.05e-22 4.87e+21 1.31e-23 3.21e-21
susinidum_oh 2.99e+02 3.34e-03 3.64e+01 2.45e+03
susinidum_coc NA NA NA NA
susinidum_pbc 5.41e+23 1.85e-24 6.03e+20 4.86e+26
susinidum_mar NA NA NA NA
psycom_dum_with 8.14e+11 1.23e-12 4.77e+10 1.39e+13
psycom_dum_study NA NA NA NA
freq_cons_dum_5day 6.27e-65 1.59e+64 5.58e-68 7.05e-62
cond_oc_dum_2inact 1.49e-18 6.70e+17 2.84e-20 7.85e-17
cond_oc_dum_3unemp NA NA NA NA
susprindum_oh 1.23e+32 8.13e-33 2.01e+29 7.53e+34
susprindum_coc 1.89e+60 5.28e-61 1.43e+57 2.51e+63
susprindum_pbc NA NA NA NA
susprindum_mar NA NA NA NA
Concordance= 1 (se = 0 )
Likelihood ratio test= 30.2 on 9 df, p=4e-04
Wald test = 1857 on 9 df, p=<2e-16
Score (logrank) test = 21.2 on 9 df, p=0.01, Robust = 9.99 p=0.4
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.00e+00 0.00e+00 8.00e+01 1.71e+24 2.95e+13 1.71e+25 Code
# Warning messages:
# 1: In Surv(time.lag, time, event) :
# Stop time must be > start time, NA created
# 2: In Surv(time.lag, time, event) :
# Stop time must be > start time, NA created
# 3: In Surv(time.lag, time, event) :
# Stop time must be > start time, NA created
# 4: In Surv(time.lag, time, event) :
# Stop time must be > start time, NA created
# 5: In Surv(time.lag, time, event) :
# Stop time must be > start time, NA created
# 6: In agreg.fit(X, Y, istrat, offset, init, control, weights = weights, :
# Loglik converged before variable 1,2,3,4,5 ; beta may be infinite.
# 7: In agreg.fit(X, Y, istrat, offset, init, control, weights = weights, :
# Ran out of iterations and did not converge
# 8: In agreg.fit(X, Y, istrat, offset, init, control, weights = weights, :
# Loglik converged before variable 17,18,19,20 ; beta may be infinite.
# 9: In agreg.fit(X, Y, istrat, offset, init, control, weights = weights, :
# Ran out of iterations and did not converge
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:
iiws_strat_alt <- list()
comb_strata_alt <- expand.grid(unique(data_mine_miss_restr_proc2$tipo_de_plan_2_mod),
unique(data_mine_miss_restr_proc2$time_interval3_alt))
for (ff in seq_along(1:nrow(comb_strata_alt))) {
i<- comb_strata_alt[ff,"Var1"]
j<- comb_strata_alt[ff,"Var2"]
data_mine_miss_proc2_subset<-
subset(data_mine_miss_restr_proc2, subset= tipo_de_plan_2_mod==i & time_interval3_alt==j)
maxfu_df_subset <- subset(maxfu_restr_df, subset= maxfu.id %in% unique(data_mine_miss_proc2_subset$id))
iiw_model_strat_alt<-
IrregLong::iiw.weights(Surv(time.lag,time,event)~
cluster(id)+ #If a frailty model is used, the cluster(id) term should appear before other covariates
tr_outcome.lag+
less_90d_tr1.lag+
log_dias_treat_imp_sin_na.lag +
comp_bpsc_y3_severe.lag+
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= data_mine_miss_proc2_subset,
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= maxfu_df_subset,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_oh", "susinidum_coc", "susinidum_pbc", "susinidum_mar", "psycom_dum_with", "psycom_dum_study_rec2", "freq_cons_dum_5day", "cond_oc_dum_3unemp", "cond_oc_dum_2inact", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "tr_outcome","log_dias_treat_imp_sin_na", "less_90d_tr1","comp_bpsc_y3_severe", "policonsumo2"),
lagfirst= c(2.95082,0,4.499811/2,0,0,0), #90/30.5 4.499811 es 90 días
first= T
)
iiws_strat_alt[[ff]] <- list(
model = iiw_model_strat_alt,
comb = paste0("Modality: ",i,"\nTime interval: ", j,"."),
cox_iiw = summary(iiw_model_strat_alt$m),
iiw = summary(iiw_model_strat_alt$iiw)
)
print(paste0("Modality: ",i,"\nTime interval: ", j))
print(summary(iiw_model_strat_alt$m))
print(summary(iiw_model_strat_alt$iiw))
}[1] "Modality: WO residential\nTime interval: [0,10]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 613, number of events= 613
(1197 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.2763 0.7586 0.2286 0.1902 -1.45
less_90d_tr1.lag -0.0115 0.9886 0.2375 0.2083 -0.06
log_dias_treat_imp_sin_na.lag -0.2477 0.7806 0.0793 0.0698 -3.55
comp_bpsc_y3_severe.lag 0.3801 1.4625 0.1943 0.1801 2.11
policonsumo2.lag -0.2659 0.7665 0.1877 0.1779 -1.49
edad_al_ing_1 -0.0252 0.9751 0.0169 0.0164 -1.54
ano_nac_corr -0.0215 0.9788 0.0163 0.0159 -1.35
susinidum_oh -0.0912 0.9128 0.2942 0.2869 -0.32
susinidum_coc -0.4372 0.6459 0.3413 0.3259 -1.34
susinidum_pbc 0.2467 1.2797 0.3006 0.3001 0.82
susinidum_mar 0.1396 1.1498 0.2900 0.2825 0.49
psycom_dum_with -0.2291 0.7952 0.1183 0.1149 -1.99
psycom_dum_study -0.0545 0.9469 0.1764 0.1707 -0.32
freq_cons_dum_5day 0.0366 1.0372 0.0990 0.0965 0.38
cond_oc_dum_2inact 0.1879 1.2068 0.1541 0.1460 1.29
cond_oc_dum_3unemp 0.1970 1.2177 0.1536 0.1476 1.33
susprindum_oh 0.4527 1.5725 0.2877 0.2560 1.77
susprindum_coc 0.5399 1.7158 0.2897 0.2441 2.21
susprindum_pbc 0.4500 1.5683 0.2700 0.2291 1.96
susprindum_mar 0.5777 1.7820 0.4281 0.4151 1.39
Pr(>|z|)
tr_outcome.lag 0.14636
less_90d_tr1.lag 0.95610
log_dias_treat_imp_sin_na.lag 0.00039 ***
comp_bpsc_y3_severe.lag 0.03478 *
policonsumo2.lag 0.13504
edad_al_ing_1 0.12342
ano_nac_corr 0.17773
susinidum_oh 0.75058
susinidum_coc 0.17976
susinidum_pbc 0.41111
susinidum_mar 0.62121
psycom_dum_with 0.04608 *
psycom_dum_study 0.74937
freq_cons_dum_5day 0.70495
cond_oc_dum_2inact 0.19799
cond_oc_dum_3unemp 0.18196
susprindum_oh 0.07699 .
susprindum_coc 0.02700 *
susprindum_pbc 0.04950 *
susprindum_mar 0.16393
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.759 1.318 0.522 1.101
less_90d_tr1.lag 0.989 1.012 0.657 1.487
log_dias_treat_imp_sin_na.lag 0.781 1.281 0.681 0.895
comp_bpsc_y3_severe.lag 1.462 0.684 1.028 2.082
policonsumo2.lag 0.766 1.305 0.541 1.086
edad_al_ing_1 0.975 1.026 0.944 1.007
ano_nac_corr 0.979 1.022 0.949 1.010
susinidum_oh 0.913 1.095 0.520 1.602
susinidum_coc 0.646 1.548 0.341 1.223
susinidum_pbc 1.280 0.781 0.711 2.304
susinidum_mar 1.150 0.870 0.661 2.000
psycom_dum_with 0.795 1.257 0.635 0.996
psycom_dum_study 0.947 1.056 0.678 1.323
freq_cons_dum_5day 1.037 0.964 0.858 1.253
cond_oc_dum_2inact 1.207 0.829 0.906 1.607
cond_oc_dum_3unemp 1.218 0.821 0.912 1.626
susprindum_oh 1.573 0.636 0.952 2.597
susprindum_coc 1.716 0.583 1.063 2.769
susprindum_pbc 1.568 0.638 1.001 2.457
susprindum_mar 1.782 0.561 0.790 4.020
Concordance= 0.608 (se = 0.012 )
Likelihood ratio test= 72.3 on 20 df, p=8e-08
Wald test = 87 on 20 df, p=2e-10
Score (logrank) test = 67.8 on 20 df, p=4e-07, Robust = 76.9 p=1e-08
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.262 0.437 0.516 0.595 0.664 2.194
[1] "Modality: GP intensive ambulatory\nTime interval: [0,10]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 3052, number of events= 3052
(5099 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.54850 0.57781 0.12658 0.11161 -4.91
less_90d_tr1.lag -0.17418 0.84015 0.10418 0.08989 -1.94
log_dias_treat_imp_sin_na.lag -0.02676 0.97359 0.03060 0.02464 -1.09
comp_bpsc_y3_severe.lag 0.11310 1.11975 0.09735 0.08271 1.37
policonsumo2.lag -0.17858 0.83646 0.11209 0.09112 -1.96
edad_al_ing_1 0.01697 1.01712 0.00806 0.00839 2.02
ano_nac_corr 0.02121 1.02144 0.00792 0.00832 2.55
susinidum_oh 0.00619 1.00621 0.13394 0.12706 0.05
susinidum_coc 0.05436 1.05587 0.15881 0.15167 0.36
susinidum_pbc 0.22952 1.25800 0.14861 0.14520 1.58
susinidum_mar 0.06074 1.06262 0.13516 0.12913 0.47
psycom_dum_with -0.06635 0.93580 0.03944 0.04029 -1.65
psycom_dum_study 0.49221 1.63593 0.06861 0.07433 6.62
freq_cons_dum_5day 0.06258 1.06458 0.03765 0.03859 1.62
cond_oc_dum_2inact 0.01356 1.01365 0.05712 0.05738 0.24
cond_oc_dum_3unemp 0.02065 1.02086 0.04043 0.04053 0.51
susprindum_oh 0.13495 1.14448 0.18044 0.18571 0.73
susprindum_coc 0.10853 1.11463 0.18058 0.18762 0.58
susprindum_pbc 0.18824 1.20712 0.17630 0.18498 1.02
susprindum_mar 0.11757 1.12476 0.19016 0.20048 0.59
Pr(>|z|)
tr_outcome.lag 8.9e-07 ***
less_90d_tr1.lag 0.053 .
log_dias_treat_imp_sin_na.lag 0.277
comp_bpsc_y3_severe.lag 0.171
policonsumo2.lag 0.050 .
edad_al_ing_1 0.043 *
ano_nac_corr 0.011 *
susinidum_oh 0.961
susinidum_coc 0.720
susinidum_pbc 0.114
susinidum_mar 0.638
psycom_dum_with 0.100 .
psycom_dum_study 3.6e-11 ***
freq_cons_dum_5day 0.105
cond_oc_dum_2inact 0.813
cond_oc_dum_3unemp 0.610
susprindum_oh 0.467
susprindum_coc 0.563
susprindum_pbc 0.309
susprindum_mar 0.558
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.578 1.731 0.464 0.719
less_90d_tr1.lag 0.840 1.190 0.704 1.002
log_dias_treat_imp_sin_na.lag 0.974 1.027 0.928 1.022
comp_bpsc_y3_severe.lag 1.120 0.893 0.952 1.317
policonsumo2.lag 0.836 1.196 0.700 1.000
edad_al_ing_1 1.017 0.983 1.001 1.034
ano_nac_corr 1.021 0.979 1.005 1.038
susinidum_oh 1.006 0.994 0.784 1.291
susinidum_coc 1.056 0.947 0.784 1.421
susinidum_pbc 1.258 0.795 0.946 1.672
susinidum_mar 1.063 0.941 0.825 1.369
psycom_dum_with 0.936 1.069 0.865 1.013
psycom_dum_study 1.636 0.611 1.414 1.893
freq_cons_dum_5day 1.065 0.939 0.987 1.148
cond_oc_dum_2inact 1.014 0.987 0.906 1.134
cond_oc_dum_3unemp 1.021 0.980 0.943 1.105
susprindum_oh 1.144 0.874 0.795 1.647
susprindum_coc 1.115 0.897 0.772 1.610
susprindum_pbc 1.207 0.828 0.840 1.735
susprindum_mar 1.125 0.889 0.759 1.666
Concordance= 0.592 (se = 0.006 )
Likelihood ratio test= 207 on 20 df, p=<2e-16
Wald test = 210 on 20 df, p=<2e-16
Score (logrank) test = 192 on 20 df, p=<2e-16, Robust = 186 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.334 0.660 0.799 0.801 0.886 2.476
[1] "Modality: basic ambulatory\nTime interval: [0,10]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 2876, number of events= 2876
(4323 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.49603 0.60894 0.13196 0.14634 -3.39
less_90d_tr1.lag -0.07117 0.93130 0.11746 0.10758 -0.66
log_dias_treat_imp_sin_na.lag -0.09772 0.90691 0.03669 0.04837 -2.02
comp_bpsc_y3_severe.lag 0.09989 1.10505 0.14866 0.12281 0.81
policonsumo2.lag -0.22168 0.80117 0.11743 0.10716 -2.07
edad_al_ing_1 0.04360 1.04456 0.00877 0.00930 4.69
ano_nac_corr 0.05061 1.05191 0.00855 0.00919 5.51
susinidum_oh 0.18366 1.20161 0.14648 0.20612 0.89
susinidum_coc 0.36574 1.44158 0.16677 0.22316 1.64
susinidum_pbc 0.26939 1.30917 0.16351 0.21714 1.24
susinidum_mar 0.28065 1.32399 0.14722 0.20772 1.35
psycom_dum_with -0.05230 0.94904 0.04144 0.04240 -1.23
psycom_dum_study 0.54576 1.72591 0.06099 0.06806 8.02
freq_cons_dum_5day 0.07322 1.07597 0.04307 0.04502 1.63
cond_oc_dum_2inact 0.09844 1.10345 0.06205 0.06203 1.59
cond_oc_dum_3unemp 0.04895 1.05017 0.04549 0.04765 1.03
susprindum_oh 0.46466 1.59147 0.22518 0.21199 2.19
susprindum_coc 0.55034 1.73384 0.22608 0.21374 2.57
susprindum_pbc 0.48029 1.61655 0.22343 0.21187 2.27
susprindum_mar 0.42294 1.52644 0.23245 0.21795 1.94
Pr(>|z|)
tr_outcome.lag 0.0007 ***
less_90d_tr1.lag 0.5083
log_dias_treat_imp_sin_na.lag 0.0434 *
comp_bpsc_y3_severe.lag 0.4160
policonsumo2.lag 0.0386 *
edad_al_ing_1 2.8e-06 ***
ano_nac_corr 3.6e-08 ***
susinidum_oh 0.3729
susinidum_coc 0.1012
susinidum_pbc 0.2147
susinidum_mar 0.1767
psycom_dum_with 0.2174
psycom_dum_study 1.1e-15 ***
freq_cons_dum_5day 0.1039
cond_oc_dum_2inact 0.1125
cond_oc_dum_3unemp 0.3043
susprindum_oh 0.0284 *
susprindum_coc 0.0100 *
susprindum_pbc 0.0234 *
susprindum_mar 0.0523 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.609 1.642 0.457 0.811
less_90d_tr1.lag 0.931 1.074 0.754 1.150
log_dias_treat_imp_sin_na.lag 0.907 1.103 0.825 0.997
comp_bpsc_y3_severe.lag 1.105 0.905 0.869 1.406
policonsumo2.lag 0.801 1.248 0.649 0.988
edad_al_ing_1 1.045 0.957 1.026 1.064
ano_nac_corr 1.052 0.951 1.033 1.071
susinidum_oh 1.202 0.832 0.802 1.800
susinidum_coc 1.442 0.694 0.931 2.233
susinidum_pbc 1.309 0.764 0.855 2.004
susinidum_mar 1.324 0.755 0.881 1.989
psycom_dum_with 0.949 1.054 0.873 1.031
psycom_dum_study 1.726 0.579 1.510 1.972
freq_cons_dum_5day 1.076 0.929 0.985 1.175
cond_oc_dum_2inact 1.103 0.906 0.977 1.246
cond_oc_dum_3unemp 1.050 0.952 0.957 1.153
susprindum_oh 1.591 0.628 1.050 2.411
susprindum_coc 1.734 0.577 1.140 2.636
susprindum_pbc 1.617 0.619 1.067 2.449
susprindum_mar 1.526 0.655 0.996 2.340
Concordance= 0.602 (se = 0.006 )
Likelihood ratio test= 253 on 20 df, p=<2e-16
Wald test = 242 on 20 df, p=<2e-16
Score (logrank) test = 238 on 20 df, p=<2e-16, Robust = 223 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.153 0.362 0.456 0.468 0.526 1.746
[1] "Modality: GP residential\nTime interval: [0,10]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1378, number of events= 1378
(2346 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.02212 0.97812 0.18230 0.14719 -0.15
less_90d_tr1.lag -0.12076 0.88624 0.14750 0.10850 -1.11
log_dias_treat_imp_sin_na.lag -0.08004 0.92308 0.04426 0.03298 -2.43
comp_bpsc_y3_severe.lag 0.00436 1.00437 0.13899 0.10793 0.04
policonsumo2.lag -0.22270 0.80035 0.15110 0.12180 -1.83
edad_al_ing_1 0.01512 1.01523 0.01204 0.01148 1.32
ano_nac_corr 0.02049 1.02070 0.01183 0.01162 1.76
susinidum_oh 0.36059 1.43418 0.21840 0.19332 1.87
susinidum_coc 0.34462 1.41145 0.25671 0.23968 1.44
susinidum_pbc 0.33183 1.39352 0.23049 0.20635 1.61
susinidum_mar 0.40104 1.49338 0.21946 0.19405 2.07
psycom_dum_with -0.07960 0.92349 0.06169 0.06166 -1.29
psycom_dum_study 0.24325 1.27539 0.10323 0.09260 2.63
freq_cons_dum_5day 0.15988 1.17337 0.06363 0.06289 2.54
cond_oc_dum_2inact 0.01093 1.01099 0.10212 0.10164 0.11
cond_oc_dum_3unemp -0.01225 0.98782 0.08671 0.08836 -0.14
susprindum_oh 0.23921 1.27025 0.29055 0.17108 1.40
susprindum_coc 0.16471 1.17905 0.29258 0.17489 0.94
susprindum_pbc 0.30173 1.35220 0.28413 0.16397 1.84
susprindum_mar 0.20262 1.22461 0.33503 0.22717 0.89
Pr(>|z|)
tr_outcome.lag 0.8805
less_90d_tr1.lag 0.2657
log_dias_treat_imp_sin_na.lag 0.0152 *
comp_bpsc_y3_severe.lag 0.9677
policonsumo2.lag 0.0675 .
edad_al_ing_1 0.1880
ano_nac_corr 0.0779 .
susinidum_oh 0.0621 .
susinidum_coc 0.1505
susinidum_pbc 0.1078
susinidum_mar 0.0388 *
psycom_dum_with 0.1967
psycom_dum_study 0.0086 **
freq_cons_dum_5day 0.0110 *
cond_oc_dum_2inact 0.9144
cond_oc_dum_3unemp 0.8897
susprindum_oh 0.1620
susprindum_coc 0.3463
susprindum_pbc 0.0657 .
susprindum_mar 0.3724
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.978 1.022 0.733 1.305
less_90d_tr1.lag 0.886 1.128 0.716 1.096
log_dias_treat_imp_sin_na.lag 0.923 1.083 0.865 0.985
comp_bpsc_y3_severe.lag 1.004 0.996 0.813 1.241
policonsumo2.lag 0.800 1.249 0.630 1.016
edad_al_ing_1 1.015 0.985 0.993 1.038
ano_nac_corr 1.021 0.980 0.998 1.044
susinidum_oh 1.434 0.697 0.982 2.095
susinidum_coc 1.411 0.708 0.882 2.258
susinidum_pbc 1.394 0.718 0.930 2.088
susinidum_mar 1.493 0.670 1.021 2.184
psycom_dum_with 0.923 1.083 0.818 1.042
psycom_dum_study 1.275 0.784 1.064 1.529
freq_cons_dum_5day 1.173 0.852 1.037 1.327
cond_oc_dum_2inact 1.011 0.989 0.828 1.234
cond_oc_dum_3unemp 0.988 1.012 0.831 1.175
susprindum_oh 1.270 0.787 0.908 1.776
susprindum_coc 1.179 0.848 0.837 1.661
susprindum_pbc 1.352 0.740 0.981 1.865
susprindum_mar 1.225 0.817 0.785 1.911
Concordance= 0.579 (se = 0.009 )
Likelihood ratio test= 56.6 on 20 df, p=2e-05
Wald test = 77.2 on 20 df, p=1e-08
Score (logrank) test = 54.3 on 20 df, p=5e-05, Robust = 66.7 p=6e-07
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.190 0.415 0.467 0.486 0.533 1.217
[1] "Modality: WO intensive ambulatory\nTime interval: [0,10]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 439, number of events= 439
(710 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -1.3012 0.2722 0.6078 0.2630 -4.95
less_90d_tr1.lag 0.2431 1.2752 0.3248 0.2038 1.19
log_dias_treat_imp_sin_na.lag 0.0482 1.0494 0.1964 0.0726 0.66
comp_bpsc_y3_severe.lag -0.1749 0.8395 0.2902 0.2065 -0.85
policonsumo2.lag 0.4071 1.5024 0.3289 0.2066 1.97
edad_al_ing_1 0.0314 1.0319 0.0239 0.0235 1.34
ano_nac_corr 0.0300 1.0304 0.0237 0.0237 1.26
susinidum_oh -0.0748 0.9280 0.3364 0.4122 -0.18
susinidum_coc -0.1454 0.8647 0.3979 0.4532 -0.32
susinidum_pbc -0.3397 0.7120 0.3672 0.4399 -0.77
susinidum_mar -0.1898 0.8272 0.3454 0.4156 -0.46
psycom_dum_with 0.2123 1.2365 0.1092 0.1048 2.03
psycom_dum_study 0.7697 2.1591 0.1897 0.1878 4.10
freq_cons_dum_5day 0.0433 1.0442 0.1019 0.1040 0.42
cond_oc_dum_2inact 0.1268 1.1352 0.1298 0.1225 1.03
cond_oc_dum_3unemp 0.1225 1.1304 0.1384 0.1315 0.93
susprindum_oh 0.3221 1.3800 0.3895 0.2781 1.16
susprindum_coc 0.6083 1.8374 0.3961 0.2869 2.12
susprindum_pbc 0.6333 1.8838 0.3803 0.2656 2.38
susprindum_mar 1.1292 3.0932 0.4219 0.2995 3.77
Pr(>|z|)
tr_outcome.lag 7.5e-07 ***
less_90d_tr1.lag 0.23300
log_dias_treat_imp_sin_na.lag 0.50647
comp_bpsc_y3_severe.lag 0.39697
policonsumo2.lag 0.04880 *
edad_al_ing_1 0.18121
ano_nac_corr 0.20670
susinidum_oh 0.85608
susinidum_coc 0.74838
susinidum_pbc 0.44007
susinidum_mar 0.64795
psycom_dum_with 0.04280 *
psycom_dum_study 4.2e-05 ***
freq_cons_dum_5day 0.67737
cond_oc_dum_2inact 0.30087
cond_oc_dum_3unemp 0.35156
susprindum_oh 0.24687
susprindum_coc 0.03399 *
susprindum_pbc 0.01712 *
susprindum_mar 0.00016 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.272 3.674 0.163 0.456
less_90d_tr1.lag 1.275 0.784 0.855 1.901
log_dias_treat_imp_sin_na.lag 1.049 0.953 0.910 1.210
comp_bpsc_y3_severe.lag 0.840 1.191 0.560 1.258
policonsumo2.lag 1.502 0.666 1.002 2.252
edad_al_ing_1 1.032 0.969 0.985 1.080
ano_nac_corr 1.030 0.970 0.984 1.080
susinidum_oh 0.928 1.078 0.414 2.082
susinidum_coc 0.865 1.156 0.356 2.102
susinidum_pbc 0.712 1.404 0.301 1.686
susinidum_mar 0.827 1.209 0.366 1.868
psycom_dum_with 1.237 0.809 1.007 1.518
psycom_dum_study 2.159 0.463 1.494 3.120
freq_cons_dum_5day 1.044 0.958 0.852 1.280
cond_oc_dum_2inact 1.135 0.881 0.893 1.443
cond_oc_dum_3unemp 1.130 0.885 0.873 1.463
susprindum_oh 1.380 0.725 0.800 2.380
susprindum_coc 1.837 0.544 1.047 3.224
susprindum_pbc 1.884 0.531 1.119 3.171
susprindum_mar 3.093 0.323 1.720 5.563
Concordance= 0.615 (se = 0.015 )
Likelihood ratio test= 46.8 on 20 df, p=6e-04
Wald test = 66.7 on 20 df, p=6e-07
Score (logrank) test = 44.6 on 20 df, p=0.001, Robust = 53.9 p=6e-05
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.138 0.362 0.482 0.507 0.590 1.829
[1] "Modality: WO residential\nTime interval: (10,20]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 510, number of events= 510
(460 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.02037 1.02058 0.32347 0.28663 0.07
less_90d_tr1.lag -0.05381 0.94761 0.47776 0.33189 -0.16
log_dias_treat_imp_sin_na.lag -0.04359 0.95734 0.13190 0.08040 -0.54
comp_bpsc_y3_severe.lag -0.22175 0.80112 0.36638 0.27653 -0.80
policonsumo2.lag -0.27505 0.75953 0.41429 0.24540 -1.12
edad_al_ing_1 -0.01107 0.98899 0.01968 0.01844 -0.60
ano_nac_corr -0.01499 0.98512 0.01893 0.01757 -0.85
susinidum_oh -0.28712 0.75043 0.29717 0.42526 -0.68
susinidum_coc -0.22276 0.80031 0.34277 0.45219 -0.49
susinidum_pbc -0.49527 0.60940 0.30801 0.43812 -1.13
susinidum_mar -0.29516 0.74441 0.29779 0.42509 -0.69
psycom_dum_with 0.11958 1.12702 0.14013 0.13615 0.88
psycom_dum_study 0.08277 1.08629 0.16386 0.16597 0.50
freq_cons_dum_5day 0.04525 1.04629 0.10570 0.10638 0.43
cond_oc_dum_2inact 0.09185 1.09620 0.16522 0.14958 0.61
cond_oc_dum_3unemp -0.05006 0.95117 0.16480 0.15168 -0.33
susprindum_oh 0.13345 1.14276 0.29473 0.29308 0.46
susprindum_coc -0.33084 0.71832 0.30710 0.31920 -1.04
susprindum_pbc 0.00516 1.00517 0.28530 0.29053 0.02
susprindum_mar -0.88875 0.41117 0.53601 0.39663 -2.24
Pr(>|z|)
tr_outcome.lag 0.943
less_90d_tr1.lag 0.871
log_dias_treat_imp_sin_na.lag 0.588
comp_bpsc_y3_severe.lag 0.423
policonsumo2.lag 0.262
edad_al_ing_1 0.548
ano_nac_corr 0.394
susinidum_oh 0.500
susinidum_coc 0.622
susinidum_pbc 0.258
susinidum_mar 0.487
psycom_dum_with 0.380
psycom_dum_study 0.618
freq_cons_dum_5day 0.671
cond_oc_dum_2inact 0.539
cond_oc_dum_3unemp 0.741
susprindum_oh 0.649
susprindum_coc 0.300
susprindum_pbc 0.986
susprindum_mar 0.025 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.021 0.980 0.582 1.790
less_90d_tr1.lag 0.948 1.055 0.494 1.816
log_dias_treat_imp_sin_na.lag 0.957 1.045 0.818 1.121
comp_bpsc_y3_severe.lag 0.801 1.248 0.466 1.377
policonsumo2.lag 0.760 1.317 0.470 1.229
edad_al_ing_1 0.989 1.011 0.954 1.025
ano_nac_corr 0.985 1.015 0.952 1.020
susinidum_oh 0.750 1.333 0.326 1.727
susinidum_coc 0.800 1.250 0.330 1.942
susinidum_pbc 0.609 1.641 0.258 1.438
susinidum_mar 0.744 1.343 0.324 1.713
psycom_dum_with 1.127 0.887 0.863 1.472
psycom_dum_study 1.086 0.921 0.785 1.504
freq_cons_dum_5day 1.046 0.956 0.849 1.289
cond_oc_dum_2inact 1.096 0.912 0.818 1.470
cond_oc_dum_3unemp 0.951 1.051 0.707 1.280
susprindum_oh 1.143 0.875 0.643 2.030
susprindum_coc 0.718 1.392 0.384 1.343
susprindum_pbc 1.005 0.995 0.569 1.776
susprindum_mar 0.411 2.432 0.189 0.895
Concordance= 0.564 (se = 0.014 )
Likelihood ratio test= 32.3 on 20 df, p=0.04
Wald test = 46.8 on 20 df, p=6e-04
Score (logrank) test = 30.6 on 20 df, p=0.06, Robust = 38.2 p=0.008
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.72 1.05 1.21 1.30 1.44 3.24
[1] "Modality: GP intensive ambulatory\nTime interval: (10,20]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 2228, number of events= 2228
(2101 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.22051 0.80211 0.20196 0.15221 -1.45
less_90d_tr1.lag 0.15824 1.17144 0.29932 0.20027 0.79
log_dias_treat_imp_sin_na.lag -0.17581 0.83878 0.06732 0.04711 -3.73
comp_bpsc_y3_severe.lag 0.04739 1.04853 0.18277 0.13092 0.36
policonsumo2.lag 0.02008 1.02029 0.22228 0.14928 0.13
edad_al_ing_1 -0.00154 0.99847 0.00887 0.00882 -0.17
ano_nac_corr -0.00251 0.99749 0.00859 0.00858 -0.29
susinidum_oh 0.05098 1.05230 0.15531 0.17300 0.29
susinidum_coc 0.15420 1.16673 0.18054 0.19068 0.81
susinidum_pbc 0.06130 1.06322 0.17251 0.19134 0.32
susinidum_mar 0.03883 1.03959 0.15757 0.17637 0.22
psycom_dum_with -0.01992 0.98028 0.04721 0.04696 -0.42
psycom_dum_study -0.05966 0.94209 0.06569 0.06467 -0.92
freq_cons_dum_5day -0.01222 0.98786 0.04390 0.04387 -0.28
cond_oc_dum_2inact 0.06559 1.06779 0.06486 0.06500 1.01
cond_oc_dum_3unemp 0.07510 1.07800 0.04738 0.04737 1.59
susprindum_oh 0.08632 1.09016 0.17649 0.18533 0.47
susprindum_coc 0.04682 1.04793 0.17742 0.18631 0.25
susprindum_pbc 0.07389 1.07668 0.17324 0.18203 0.41
susprindum_mar 0.13331 1.14260 0.19634 0.20113 0.66
Pr(>|z|)
tr_outcome.lag 0.14742
less_90d_tr1.lag 0.42946
log_dias_treat_imp_sin_na.lag 0.00019 ***
comp_bpsc_y3_severe.lag 0.71735
policonsumo2.lag 0.89298
edad_al_ing_1 0.86174
ano_nac_corr 0.76954
susinidum_oh 0.76823
susinidum_coc 0.41870
susinidum_pbc 0.74870
susinidum_mar 0.82576
psycom_dum_with 0.67141
psycom_dum_study 0.35627
freq_cons_dum_5day 0.78062
cond_oc_dum_2inact 0.31295
cond_oc_dum_3unemp 0.11288
susprindum_oh 0.64136
susprindum_coc 0.80159
susprindum_pbc 0.68481
susprindum_mar 0.50745
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.802 1.247 0.595 1.08
less_90d_tr1.lag 1.171 0.854 0.791 1.73
log_dias_treat_imp_sin_na.lag 0.839 1.192 0.765 0.92
comp_bpsc_y3_severe.lag 1.049 0.954 0.811 1.36
policonsumo2.lag 1.020 0.980 0.761 1.37
edad_al_ing_1 0.998 1.002 0.981 1.02
ano_nac_corr 0.997 1.003 0.981 1.01
susinidum_oh 1.052 0.950 0.750 1.48
susinidum_coc 1.167 0.857 0.803 1.70
susinidum_pbc 1.063 0.941 0.731 1.55
susinidum_mar 1.040 0.962 0.736 1.47
psycom_dum_with 0.980 1.020 0.894 1.07
psycom_dum_study 0.942 1.061 0.830 1.07
freq_cons_dum_5day 0.988 1.012 0.906 1.08
cond_oc_dum_2inact 1.068 0.937 0.940 1.21
cond_oc_dum_3unemp 1.078 0.928 0.982 1.18
susprindum_oh 1.090 0.917 0.758 1.57
susprindum_coc 1.048 0.954 0.727 1.51
susprindum_pbc 1.077 0.929 0.754 1.54
susprindum_mar 1.143 0.875 0.770 1.69
Concordance= 0.529 (se = 0.007 )
Likelihood ratio test= 68.9 on 20 df, p=3e-07
Wald test = 96.2 on 20 df, p=6e-12
Score (logrank) test = 59.1 on 20 df, p=1e-05, Robust = 49.1 p=3e-04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.589 0.812 0.847 0.886 0.877 2.101
[1] "Modality: basic ambulatory\nTime interval: (10,20]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1787, number of events= 1787
(1705 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.12932 0.87869 0.25813 0.18571 -0.70
less_90d_tr1.lag -0.15985 0.85227 0.42237 0.34156 -0.47
log_dias_treat_imp_sin_na.lag -0.12757 0.88023 0.08020 0.06162 -2.07
comp_bpsc_y3_severe.lag -0.08762 0.91610 0.28847 0.22146 -0.40
policonsumo2.lag 0.21017 1.23388 0.24199 0.20255 1.04
edad_al_ing_1 -0.00577 0.99424 0.01003 0.01001 -0.58
ano_nac_corr -0.00986 0.99019 0.00973 0.00982 -1.00
susinidum_oh -0.10656 0.89892 0.17572 0.19057 -0.56
susinidum_coc -0.25140 0.77771 0.20950 0.21642 -1.16
susinidum_pbc -0.02585 0.97448 0.19592 0.21091 -0.12
susinidum_mar -0.07484 0.92789 0.17887 0.19301 -0.39
psycom_dum_with -0.05706 0.94453 0.05251 0.05189 -1.10
psycom_dum_study -0.15100 0.85984 0.07230 0.07389 -2.04
freq_cons_dum_5day 0.02770 1.02809 0.05517 0.05588 0.50
cond_oc_dum_2inact -0.02770 0.97268 0.07139 0.07290 -0.38
cond_oc_dum_3unemp -0.01719 0.98295 0.05936 0.05787 -0.30
susprindum_oh 0.04424 1.04523 0.20165 0.17037 0.26
susprindum_coc 0.06049 1.06235 0.20551 0.17376 0.35
susprindum_pbc -0.01698 0.98316 0.20146 0.17168 -0.10
susprindum_mar 0.21749 1.24296 0.21387 0.18526 1.17
Pr(>|z|)
tr_outcome.lag 0.486
less_90d_tr1.lag 0.640
log_dias_treat_imp_sin_na.lag 0.038 *
comp_bpsc_y3_severe.lag 0.692
policonsumo2.lag 0.299
edad_al_ing_1 0.564
ano_nac_corr 0.316
susinidum_oh 0.576
susinidum_coc 0.245
susinidum_pbc 0.902
susinidum_mar 0.698
psycom_dum_with 0.271
psycom_dum_study 0.041 *
freq_cons_dum_5day 0.620
cond_oc_dum_2inact 0.704
cond_oc_dum_3unemp 0.766
susprindum_oh 0.795
susprindum_coc 0.728
susprindum_pbc 0.921
susprindum_mar 0.240
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.879 1.138 0.611 1.264
less_90d_tr1.lag 0.852 1.173 0.436 1.665
log_dias_treat_imp_sin_na.lag 0.880 1.136 0.780 0.993
comp_bpsc_y3_severe.lag 0.916 1.092 0.594 1.414
policonsumo2.lag 1.234 0.810 0.830 1.835
edad_al_ing_1 0.994 1.006 0.975 1.014
ano_nac_corr 0.990 1.010 0.971 1.009
susinidum_oh 0.899 1.112 0.619 1.306
susinidum_coc 0.778 1.286 0.509 1.189
susinidum_pbc 0.974 1.026 0.645 1.473
susinidum_mar 0.928 1.078 0.636 1.355
psycom_dum_with 0.945 1.059 0.853 1.046
psycom_dum_study 0.860 1.163 0.744 0.994
freq_cons_dum_5day 1.028 0.973 0.921 1.147
cond_oc_dum_2inact 0.973 1.028 0.843 1.122
cond_oc_dum_3unemp 0.983 1.017 0.878 1.101
susprindum_oh 1.045 0.957 0.749 1.460
susprindum_coc 1.062 0.941 0.756 1.493
susprindum_pbc 0.983 1.017 0.702 1.376
susprindum_mar 1.243 0.805 0.864 1.787
Concordance= 0.535 (se = 0.008 )
Likelihood ratio test= 30.9 on 20 df, p=0.06
Wald test = 36.3 on 20 df, p=0.01
Score (logrank) test = 29.6 on 20 df, p=0.08, Robust = 30.6 p=0.06
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.746 1.027 1.092 1.115 1.173 2.260
[1] "Modality: GP residential\nTime interval: (10,20]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1104, number of events= 1104
(1012 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.16084 1.17450 0.22786 0.16977 0.95
less_90d_tr1.lag -0.49178 0.61154 0.34685 0.23196 -2.12
log_dias_treat_imp_sin_na.lag -0.34202 0.71034 0.10550 0.06216 -5.50
comp_bpsc_y3_severe.lag 0.30136 1.35170 0.25249 0.19164 1.57
policonsumo2.lag 0.30428 1.35564 0.34606 0.18312 1.66
edad_al_ing_1 0.00602 1.00604 0.01385 0.01406 0.43
ano_nac_corr 0.00667 1.00669 0.01397 0.01379 0.48
susinidum_oh -0.15110 0.85976 0.21234 0.18102 -0.83
susinidum_coc -0.03124 0.96925 0.26184 0.22544 -0.14
susinidum_pbc -0.28296 0.75355 0.22751 0.19856 -1.43
susinidum_mar -0.13798 0.87111 0.21120 0.17930 -0.77
psycom_dum_with -0.07512 0.92763 0.07035 0.06946 -1.08
psycom_dum_study -0.19636 0.82172 0.09873 0.09277 -2.12
freq_cons_dum_5day -0.02582 0.97451 0.06870 0.06894 -0.37
cond_oc_dum_2inact -0.00504 0.99497 0.11271 0.11272 -0.04
cond_oc_dum_3unemp 0.11897 1.12634 0.09198 0.08874 1.34
susprindum_oh -0.03578 0.96485 0.31288 0.27870 -0.13
susprindum_coc -0.16252 0.85000 0.31542 0.28194 -0.58
susprindum_pbc -0.15322 0.85794 0.30652 0.27433 -0.56
susprindum_mar 0.41912 1.52062 0.35401 0.32844 1.28
Pr(>|z|)
tr_outcome.lag 0.343
less_90d_tr1.lag 0.034 *
log_dias_treat_imp_sin_na.lag 3.8e-08 ***
comp_bpsc_y3_severe.lag 0.116
policonsumo2.lag 0.097 .
edad_al_ing_1 0.669
ano_nac_corr 0.629
susinidum_oh 0.404
susinidum_coc 0.890
susinidum_pbc 0.154
susinidum_mar 0.442
psycom_dum_with 0.279
psycom_dum_study 0.034 *
freq_cons_dum_5day 0.708
cond_oc_dum_2inact 0.964
cond_oc_dum_3unemp 0.180
susprindum_oh 0.898
susprindum_coc 0.564
susprindum_pbc 0.576
susprindum_mar 0.202
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.175 0.851 0.842 1.638
less_90d_tr1.lag 0.612 1.635 0.388 0.964
log_dias_treat_imp_sin_na.lag 0.710 1.408 0.629 0.802
comp_bpsc_y3_severe.lag 1.352 0.740 0.928 1.968
policonsumo2.lag 1.356 0.738 0.947 1.941
edad_al_ing_1 1.006 0.994 0.979 1.034
ano_nac_corr 1.007 0.993 0.980 1.034
susinidum_oh 0.860 1.163 0.603 1.226
susinidum_coc 0.969 1.032 0.623 1.508
susinidum_pbc 0.754 1.327 0.511 1.112
susinidum_mar 0.871 1.148 0.613 1.238
psycom_dum_with 0.928 1.078 0.810 1.063
psycom_dum_study 0.822 1.217 0.685 0.986
freq_cons_dum_5day 0.975 1.026 0.851 1.115
cond_oc_dum_2inact 0.995 1.005 0.798 1.241
cond_oc_dum_3unemp 1.126 0.888 0.947 1.340
susprindum_oh 0.965 1.036 0.559 1.666
susprindum_coc 0.850 1.176 0.489 1.477
susprindum_pbc 0.858 1.166 0.501 1.469
susprindum_mar 1.521 0.658 0.799 2.895
Concordance= 0.551 (se = 0.01 )
Likelihood ratio test= 53.5 on 20 df, p=7e-05
Wald test = 82.3 on 20 df, p=2e-09
Score (logrank) test = 50.3 on 20 df, p=2e-04, Robust = 51.4 p=1e-04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.59 1.14 1.22 1.30 1.35 4.01
[1] "Modality: WO intensive ambulatory\nTime interval: (10,20]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 383, number of events= 383
(363 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.0139 1.0140 0.5205 0.2113 0.07
less_90d_tr1.lag -0.1059 0.8995 0.6503 0.2745 -0.39
log_dias_treat_imp_sin_na.lag -0.2009 0.8180 0.2031 0.0842 -2.38
comp_bpsc_y3_severe.lag 0.2002 1.2216 0.4707 0.2543 0.79
policonsumo2.lag 0.1453 1.1563 0.5554 0.2622 0.55
edad_al_ing_1 0.0181 1.0183 0.0229 0.0233 0.78
ano_nac_corr 0.0110 1.0110 0.0226 0.0218 0.50
susinidum_oh -0.1997 0.8190 0.3295 0.2298 -0.87
susinidum_coc -0.2145 0.8069 0.3881 0.2883 -0.74
susinidum_pbc -0.3852 0.6803 0.3518 0.2681 -1.44
susinidum_mar -0.2611 0.7702 0.3303 0.2274 -1.15
psycom_dum_with -0.0695 0.9329 0.1233 0.1172 -0.59
psycom_dum_study -0.0870 0.9166 0.1661 0.1586 -0.55
freq_cons_dum_5day 0.0223 1.0226 0.1079 0.1065 0.21
cond_oc_dum_2inact -0.2520 0.7772 0.1402 0.1450 -1.74
cond_oc_dum_3unemp -0.1516 0.8594 0.1437 0.1478 -1.03
susprindum_oh -0.1029 0.9022 0.4157 0.3344 -0.31
susprindum_coc 0.1311 1.1400 0.4138 0.3210 0.41
susprindum_pbc 0.0489 1.0501 0.4034 0.3065 0.16
susprindum_mar 0.3545 1.4254 0.4304 0.3513 1.01
Pr(>|z|)
tr_outcome.lag 0.948
less_90d_tr1.lag 0.700
log_dias_treat_imp_sin_na.lag 0.017 *
comp_bpsc_y3_severe.lag 0.431
policonsumo2.lag 0.580
edad_al_ing_1 0.435
ano_nac_corr 0.615
susinidum_oh 0.385
susinidum_coc 0.457
susinidum_pbc 0.151
susinidum_mar 0.251
psycom_dum_with 0.553
psycom_dum_study 0.583
freq_cons_dum_5day 0.834
cond_oc_dum_2inact 0.082 .
cond_oc_dum_3unemp 0.305
susprindum_oh 0.758
susprindum_coc 0.683
susprindum_pbc 0.873
susprindum_mar 0.313
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.014 0.986 0.670 1.534
less_90d_tr1.lag 0.900 1.112 0.525 1.541
log_dias_treat_imp_sin_na.lag 0.818 1.222 0.694 0.965
comp_bpsc_y3_severe.lag 1.222 0.819 0.742 2.011
policonsumo2.lag 1.156 0.865 0.692 1.933
edad_al_ing_1 1.018 0.982 0.973 1.066
ano_nac_corr 1.011 0.989 0.969 1.055
susinidum_oh 0.819 1.221 0.522 1.285
susinidum_coc 0.807 1.239 0.459 1.420
susinidum_pbc 0.680 1.470 0.402 1.150
susinidum_mar 0.770 1.298 0.493 1.203
psycom_dum_with 0.933 1.072 0.741 1.174
psycom_dum_study 0.917 1.091 0.672 1.251
freq_cons_dum_5day 1.023 0.978 0.830 1.260
cond_oc_dum_2inact 0.777 1.287 0.585 1.033
cond_oc_dum_3unemp 0.859 1.164 0.643 1.148
susprindum_oh 0.902 1.108 0.468 1.738
susprindum_coc 1.140 0.877 0.608 2.139
susprindum_pbc 1.050 0.952 0.576 1.915
susprindum_mar 1.425 0.702 0.716 2.838
Concordance= 0.56 (se = 0.016 )
Likelihood ratio test= 14.7 on 20 df, p=0.8
Wald test = 23.1 on 20 df, p=0.3
Score (logrank) test = 15 on 20 df, p=0.8, Robust = 21.1 p=0.4
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.73 1.28 1.46 1.47 1.65 2.79
[1] "Modality: WO residential\nTime interval: (20,30]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 337, number of events= 337
(312 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.52775 0.58993 0.59758 0.55588 -0.95
less_90d_tr1.lag 1.28742 3.62343 0.63270 0.69793 1.84
log_dias_treat_imp_sin_na.lag 0.37313 1.45227 0.21591 0.17707 2.11
comp_bpsc_y3_severe.lag -0.44744 0.63926 0.53809 0.42889 -1.04
policonsumo2.lag -1.35950 0.25679 0.63831 0.50568 -2.69
edad_al_ing_1 0.00477 1.00478 0.02813 0.02935 0.16
ano_nac_corr 0.01147 1.01153 0.02742 0.02843 0.40
susinidum_oh -0.90099 0.40617 0.36726 0.34401 -2.62
susinidum_coc -0.42751 0.65213 0.44421 0.46878 -0.91
susinidum_pbc -0.80738 0.44602 0.38152 0.35273 -2.29
susinidum_mar -0.67932 0.50696 0.37085 0.34238 -1.98
psycom_dum_with 0.10540 1.11116 0.16856 0.16076 0.66
psycom_dum_study 0.16165 1.17544 0.19683 0.18720 0.86
freq_cons_dum_5day -0.11891 0.88789 0.13535 0.12817 -0.93
cond_oc_dum_2inact 0.01787 1.01803 0.21311 0.20359 0.09
cond_oc_dum_3unemp 0.11626 1.12328 0.21418 0.20141 0.58
susprindum_oh 0.60703 1.83497 0.36513 0.33605 1.81
susprindum_coc 0.19494 1.21524 0.36571 0.34826 0.56
susprindum_pbc 0.54841 1.73050 0.34360 0.32407 1.69
susprindum_mar -0.07329 0.92933 0.49572 0.46633 -0.16
Pr(>|z|)
tr_outcome.lag 0.3424
less_90d_tr1.lag 0.0651 .
log_dias_treat_imp_sin_na.lag 0.0351 *
comp_bpsc_y3_severe.lag 0.2968
policonsumo2.lag 0.0072 **
edad_al_ing_1 0.8709
ano_nac_corr 0.6868
susinidum_oh 0.0088 **
susinidum_coc 0.3618
susinidum_pbc 0.0221 *
susinidum_mar 0.0472 *
psycom_dum_with 0.5120
psycom_dum_study 0.3879
freq_cons_dum_5day 0.3535
cond_oc_dum_2inact 0.9301
cond_oc_dum_3unemp 0.5638
susprindum_oh 0.0709 .
susprindum_coc 0.5756
susprindum_pbc 0.0906 .
susprindum_mar 0.8751
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.590 1.695 0.1984 1.754
less_90d_tr1.lag 3.623 0.276 0.9227 14.230
log_dias_treat_imp_sin_na.lag 1.452 0.689 1.0264 2.055
comp_bpsc_y3_severe.lag 0.639 1.564 0.2758 1.482
policonsumo2.lag 0.257 3.894 0.0953 0.692
edad_al_ing_1 1.005 0.995 0.9486 1.064
ano_nac_corr 1.012 0.989 0.9567 1.070
susinidum_oh 0.406 2.462 0.2070 0.797
susinidum_coc 0.652 1.533 0.2602 1.634
susinidum_pbc 0.446 2.242 0.2234 0.890
susinidum_mar 0.507 1.973 0.2591 0.992
psycom_dum_with 1.111 0.900 0.8109 1.523
psycom_dum_study 1.175 0.851 0.8144 1.696
freq_cons_dum_5day 0.888 1.126 0.6907 1.141
cond_oc_dum_2inact 1.018 0.982 0.6831 1.517
cond_oc_dum_3unemp 1.123 0.890 0.7569 1.667
susprindum_oh 1.835 0.545 0.9497 3.545
susprindum_coc 1.215 0.823 0.6141 2.405
susprindum_pbc 1.731 0.578 0.9169 3.266
susprindum_mar 0.929 1.076 0.3726 2.318
Concordance= 0.58 (se = 0.019 )
Likelihood ratio test= 34 on 20 df, p=0.03
Wald test = 30.1 on 20 df, p=0.07
Score (logrank) test = 32.6 on 20 df, p=0.04, Robust = 30.6 p=0.06
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.49 1.15 1.33 1.46 1.54 5.65
[1] "Modality: GP intensive ambulatory\nTime interval: (20,30]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1468, number of events= 1468
(1416 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.77020 0.46292 0.32283 0.21267 -3.62
less_90d_tr1.lag 0.75061 2.11829 0.49073 0.32418 2.32
log_dias_treat_imp_sin_na.lag -0.01781 0.98235 0.12088 0.06905 -0.26
comp_bpsc_y3_severe.lag 0.13237 1.14153 0.27442 0.18921 0.70
policonsumo2.lag -0.09052 0.91346 0.39153 0.19891 -0.46
edad_al_ing_1 0.00928 1.00932 0.01199 0.01157 0.80
ano_nac_corr 0.01265 1.01273 0.01185 0.01153 1.10
susinidum_oh -0.17332 0.84087 0.17260 0.14459 -1.20
susinidum_coc -0.12314 0.88414 0.20924 0.18679 -0.66
susinidum_pbc -0.24649 0.78154 0.19359 0.16674 -1.48
susinidum_mar -0.14682 0.86345 0.17517 0.14808 -0.99
psycom_dum_with 0.07150 1.07412 0.05853 0.05778 1.24
psycom_dum_study 0.13922 1.14937 0.07958 0.08264 1.68
freq_cons_dum_5day -0.02035 0.97985 0.05446 0.05442 -0.37
cond_oc_dum_2inact -0.16760 0.84569 0.08044 0.07567 -2.21
cond_oc_dum_3unemp -0.07804 0.92492 0.05854 0.05876 -1.33
susprindum_oh 0.34743 1.41543 0.22476 0.22378 1.55
susprindum_coc 0.37297 1.45205 0.22768 0.22798 1.64
susprindum_pbc 0.36533 1.44099 0.21944 0.21990 1.66
susprindum_mar 0.54446 1.72368 0.24308 0.24195 2.25
Pr(>|z|)
tr_outcome.lag 0.00029 ***
less_90d_tr1.lag 0.02059 *
log_dias_treat_imp_sin_na.lag 0.79649
comp_bpsc_y3_severe.lag 0.48419
policonsumo2.lag 0.64906
edad_al_ing_1 0.42279
ano_nac_corr 0.27261
susinidum_oh 0.23063
susinidum_coc 0.50974
susinidum_pbc 0.13932
susinidum_mar 0.32145
psycom_dum_with 0.21597
psycom_dum_study 0.09206 .
freq_cons_dum_5day 0.70838
cond_oc_dum_2inact 0.02676 *
cond_oc_dum_3unemp 0.18414
susprindum_oh 0.12052
susprindum_coc 0.10184
susprindum_pbc 0.09665 .
susprindum_mar 0.02443 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.463 2.160 0.305 0.702
less_90d_tr1.lag 2.118 0.472 1.122 3.999
log_dias_treat_imp_sin_na.lag 0.982 1.018 0.858 1.125
comp_bpsc_y3_severe.lag 1.142 0.876 0.788 1.654
policonsumo2.lag 0.913 1.095 0.619 1.349
edad_al_ing_1 1.009 0.991 0.987 1.032
ano_nac_corr 1.013 0.987 0.990 1.036
susinidum_oh 0.841 1.189 0.633 1.116
susinidum_coc 0.884 1.131 0.613 1.275
susinidum_pbc 0.782 1.280 0.564 1.084
susinidum_mar 0.863 1.158 0.646 1.154
psycom_dum_with 1.074 0.931 0.959 1.203
psycom_dum_study 1.149 0.870 0.978 1.351
freq_cons_dum_5day 0.980 1.021 0.881 1.090
cond_oc_dum_2inact 0.846 1.182 0.729 0.981
cond_oc_dum_3unemp 0.925 1.081 0.824 1.038
susprindum_oh 1.415 0.706 0.913 2.195
susprindum_coc 1.452 0.689 0.929 2.270
susprindum_pbc 1.441 0.694 0.936 2.217
susprindum_mar 1.724 0.580 1.073 2.770
Concordance= 0.537 (se = 0.009 )
Likelihood ratio test= 45.9 on 20 df, p=8e-04
Wald test = 52 on 20 df, p=1e-04
Score (logrank) test = 41.3 on 20 df, p=0.003, Robust = 34.9 p=0.02
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.513 0.767 0.823 0.850 0.890 2.288
[1] "Modality: basic ambulatory\nTime interval: (20,30]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1208, number of events= 1208
(1172 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.05575 1.05733 0.41651 0.26438 0.21
less_90d_tr1.lag -0.45329 0.63554 0.43659 0.35253 -1.29
log_dias_treat_imp_sin_na.lag -0.15252 0.85854 0.14076 0.08064 -1.89
comp_bpsc_y3_severe.lag -0.55325 0.57508 0.40430 0.31080 -1.78
policonsumo2.lag 0.25287 1.28771 0.43978 0.28250 0.90
edad_al_ing_1 0.04213 1.04303 0.01381 0.01319 3.19
ano_nac_corr 0.04128 1.04214 0.01367 0.01295 3.19
susinidum_oh 0.01952 1.01971 0.21963 0.21984 0.09
susinidum_coc 0.09272 1.09716 0.25589 0.25100 0.37
susinidum_pbc 0.21430 1.23900 0.24964 0.24453 0.88
susinidum_mar 0.05777 1.05947 0.22158 0.22321 0.26
psycom_dum_with -0.05744 0.94418 0.06662 0.06799 -0.84
psycom_dum_study -0.10343 0.90174 0.07985 0.07481 -1.38
freq_cons_dum_5day 0.07591 1.07887 0.06736 0.06549 1.16
cond_oc_dum_2inact -0.00385 0.99616 0.09586 0.09454 -0.04
cond_oc_dum_3unemp -0.02597 0.97437 0.06951 0.07100 -0.37
susprindum_oh -0.44138 0.64315 0.28577 0.22911 -1.93
susprindum_coc -0.44717 0.63943 0.29004 0.23458 -1.91
susprindum_pbc -0.47769 0.62021 0.28560 0.22746 -2.10
susprindum_mar -0.44252 0.64242 0.29882 0.25360 -1.74
Pr(>|z|)
tr_outcome.lag 0.8330
less_90d_tr1.lag 0.1985
log_dias_treat_imp_sin_na.lag 0.0586 .
comp_bpsc_y3_severe.lag 0.0751 .
policonsumo2.lag 0.3707
edad_al_ing_1 0.0014 **
ano_nac_corr 0.0014 **
susinidum_oh 0.9292
susinidum_coc 0.7118
susinidum_pbc 0.3808
susinidum_mar 0.7958
psycom_dum_with 0.3983
psycom_dum_study 0.1668
freq_cons_dum_5day 0.2464
cond_oc_dum_2inact 0.9675
cond_oc_dum_3unemp 0.7146
susprindum_oh 0.0540 .
susprindum_coc 0.0566 .
susprindum_pbc 0.0357 *
susprindum_mar 0.0810 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.057 0.946 0.630 1.775
less_90d_tr1.lag 0.636 1.573 0.318 1.268
log_dias_treat_imp_sin_na.lag 0.859 1.165 0.733 1.006
comp_bpsc_y3_severe.lag 0.575 1.739 0.313 1.057
policonsumo2.lag 1.288 0.777 0.740 2.240
edad_al_ing_1 1.043 0.959 1.016 1.070
ano_nac_corr 1.042 0.960 1.016 1.069
susinidum_oh 1.020 0.981 0.663 1.569
susinidum_coc 1.097 0.911 0.671 1.794
susinidum_pbc 1.239 0.807 0.767 2.001
susinidum_mar 1.059 0.944 0.684 1.641
psycom_dum_with 0.944 1.059 0.826 1.079
psycom_dum_study 0.902 1.109 0.779 1.044
freq_cons_dum_5day 1.079 0.927 0.949 1.227
cond_oc_dum_2inact 0.996 1.004 0.828 1.199
cond_oc_dum_3unemp 0.974 1.026 0.848 1.120
susprindum_oh 0.643 1.555 0.410 1.008
susprindum_coc 0.639 1.564 0.404 1.013
susprindum_pbc 0.620 1.612 0.397 0.969
susprindum_mar 0.642 1.557 0.391 1.056
Concordance= 0.553 (se = 0.009 )
Likelihood ratio test= 30.9 on 20 df, p=0.06
Wald test = 42.6 on 20 df, p=0.002
Score (logrank) test = 29.1 on 20 df, p=0.09, Robust = 31.7 p=0.05
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.89 1.40 1.54 1.56 1.67 4.05
[1] "Modality: GP residential\nTime interval: (20,30]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 612, number of events= 612
(578 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.1769 0.8378 0.4834 0.2907 -0.61
less_90d_tr1.lag -0.2906 0.7478 0.5881 0.3684 -0.79
log_dias_treat_imp_sin_na.lag -0.3352 0.7152 0.1787 0.1274 -2.63
comp_bpsc_y3_severe.lag 0.3234 1.3818 0.4523 0.3102 1.04
policonsumo2.lag 0.4513 1.5704 0.5524 0.3767 1.20
edad_al_ing_1 0.0155 1.0157 0.0203 0.0182 0.85
ano_nac_corr 0.0204 1.0206 0.0200 0.0183 1.11
susinidum_oh -0.4689 0.6257 0.3172 0.2629 -1.78
susinidum_coc -0.0629 0.9391 0.3628 0.3023 -0.21
susinidum_pbc -0.3090 0.7342 0.3327 0.2733 -1.13
susinidum_mar -0.4797 0.6190 0.3174 0.2605 -1.84
psycom_dum_with -0.1006 0.9043 0.0943 0.0977 -1.03
psycom_dum_study -0.0192 0.9810 0.1339 0.1244 -0.15
freq_cons_dum_5day 0.1309 1.1398 0.0948 0.0945 1.38
cond_oc_dum_2inact -0.0699 0.9325 0.1497 0.1480 -0.47
cond_oc_dum_3unemp -0.0659 0.9363 0.1298 0.1280 -0.51
susprindum_oh 0.8377 2.3110 0.4654 0.4005 2.09
susprindum_coc 0.8259 2.2840 0.4701 0.4049 2.04
susprindum_pbc 0.8141 2.2571 0.4569 0.3882 2.10
susprindum_mar 0.5328 1.7036 0.5248 0.4425 1.20
Pr(>|z|)
tr_outcome.lag 0.5428
less_90d_tr1.lag 0.4302
log_dias_treat_imp_sin_na.lag 0.0085 **
comp_bpsc_y3_severe.lag 0.2971
policonsumo2.lag 0.2308
edad_al_ing_1 0.3945
ano_nac_corr 0.2652
susinidum_oh 0.0745 .
susinidum_coc 0.8352
susinidum_pbc 0.2582
susinidum_mar 0.0655 .
psycom_dum_with 0.3033
psycom_dum_study 0.8772
freq_cons_dum_5day 0.1661
cond_oc_dum_2inact 0.6369
cond_oc_dum_3unemp 0.6069
susprindum_oh 0.0365 *
susprindum_coc 0.0413 *
susprindum_pbc 0.0360 *
susprindum_mar 0.2286
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.838 1.194 0.474 1.481
less_90d_tr1.lag 0.748 1.337 0.363 1.540
log_dias_treat_imp_sin_na.lag 0.715 1.398 0.557 0.918
comp_bpsc_y3_severe.lag 1.382 0.724 0.752 2.538
policonsumo2.lag 1.570 0.637 0.751 3.286
edad_al_ing_1 1.016 0.985 0.980 1.053
ano_nac_corr 1.021 0.980 0.985 1.058
susinidum_oh 0.626 1.598 0.374 1.048
susinidum_coc 0.939 1.065 0.519 1.698
susinidum_pbc 0.734 1.362 0.430 1.254
susinidum_mar 0.619 1.616 0.372 1.031
psycom_dum_with 0.904 1.106 0.747 1.095
psycom_dum_study 0.981 1.019 0.769 1.252
freq_cons_dum_5day 1.140 0.877 0.947 1.372
cond_oc_dum_2inact 0.933 1.072 0.698 1.246
cond_oc_dum_3unemp 0.936 1.068 0.729 1.203
susprindum_oh 2.311 0.433 1.054 5.066
susprindum_coc 2.284 0.438 1.033 5.051
susprindum_pbc 2.257 0.443 1.055 4.830
susprindum_mar 1.704 0.587 0.716 4.055
Concordance= 0.549 (se = 0.013 )
Likelihood ratio test= 30.6 on 20 df, p=0.06
Wald test = 49.6 on 20 df, p=3e-04
Score (logrank) test = 28.9 on 20 df, p=0.09, Robust = 32.6 p=0.04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.287 0.618 0.674 0.699 0.734 2.114
[1] "Modality: WO intensive ambulatory\nTime interval: (20,30]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 247, number of events= 247
(235 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -7.22e-02 9.30e-01 8.60e-01 4.46e-01 -0.16
less_90d_tr1.lag -6.96e-01 4.98e-01 9.59e-01 5.93e-01 -1.17
log_dias_treat_imp_sin_na.lag -4.77e-01 6.21e-01 3.69e-01 2.08e-01 -2.29
comp_bpsc_y3_severe.lag 4.74e-01 1.61e+00 8.48e-01 5.41e-01 0.88
policonsumo2.lag 9.90e-01 2.69e+00 8.05e-01 5.92e-01 1.67
edad_al_ing_1 -1.56e-02 9.85e-01 3.12e-02 2.97e-02 -0.53
ano_nac_corr -1.85e-02 9.82e-01 2.94e-02 2.69e-02 -0.69
susinidum_oh -5.46e-01 5.79e-01 5.57e-01 4.73e-01 -1.15
susinidum_coc -4.26e-01 6.53e-01 6.33e-01 5.47e-01 -0.78
susinidum_pbc -7.19e-01 4.87e-01 5.93e-01 5.02e-01 -1.43
susinidum_mar -4.38e-01 6.45e-01 5.65e-01 4.76e-01 -0.92
psycom_dum_with 2.34e-02 1.02e+00 1.46e-01 1.40e-01 0.17
psycom_dum_study 2.92e-01 1.34e+00 2.35e-01 2.58e-01 1.13
freq_cons_dum_5day -2.47e-01 7.81e-01 1.47e-01 1.42e-01 -1.74
cond_oc_dum_2inact -6.83e-02 9.34e-01 1.74e-01 1.56e-01 -0.44
cond_oc_dum_3unemp -2.42e-02 9.76e-01 1.84e-01 1.81e-01 -0.13
susprindum_oh -4.26e-01 6.53e-01 4.59e-01 3.61e-01 -1.18
susprindum_coc 1.33e-01 1.14e+00 4.72e-01 3.97e-01 0.33
susprindum_pbc 3.51e-05 1.00e+00 4.63e-01 3.79e-01 0.00
susprindum_mar -1.33e-02 9.87e-01 5.19e-01 4.06e-01 -0.03
Pr(>|z|)
tr_outcome.lag 0.872
less_90d_tr1.lag 0.241
log_dias_treat_imp_sin_na.lag 0.022 *
comp_bpsc_y3_severe.lag 0.381
policonsumo2.lag 0.095 .
edad_al_ing_1 0.599
ano_nac_corr 0.493
susinidum_oh 0.248
susinidum_coc 0.436
susinidum_pbc 0.152
susinidum_mar 0.357
psycom_dum_with 0.867
psycom_dum_study 0.257
freq_cons_dum_5day 0.082 .
cond_oc_dum_2inact 0.662
cond_oc_dum_3unemp 0.894
susprindum_oh 0.238
susprindum_coc 0.739
susprindum_pbc 1.000
susprindum_mar 0.974
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.930 1.075 0.388 2.232
less_90d_tr1.lag 0.498 2.006 0.156 1.595
log_dias_treat_imp_sin_na.lag 0.621 1.611 0.413 0.933
comp_bpsc_y3_severe.lag 1.607 0.622 0.556 4.640
policonsumo2.lag 2.691 0.372 0.843 8.589
edad_al_ing_1 0.985 1.016 0.929 1.044
ano_nac_corr 0.982 1.019 0.931 1.035
susinidum_oh 0.579 1.727 0.229 1.464
susinidum_coc 0.653 1.532 0.223 1.907
susinidum_pbc 0.487 2.052 0.182 1.303
susinidum_mar 0.645 1.550 0.254 1.639
psycom_dum_with 1.024 0.977 0.778 1.347
psycom_dum_study 1.339 0.747 0.808 2.219
freq_cons_dum_5day 0.781 1.281 0.591 1.032
cond_oc_dum_2inact 0.934 1.071 0.687 1.269
cond_oc_dum_3unemp 0.976 1.024 0.684 1.392
susprindum_oh 0.653 1.531 0.322 1.325
susprindum_coc 1.142 0.876 0.524 2.487
susprindum_pbc 1.000 1.000 0.476 2.102
susprindum_mar 0.987 1.013 0.445 2.187
Concordance= 0.587 (se = 0.021 )
Likelihood ratio test= 17.1 on 20 df, p=0.6
Wald test = 23.7 on 20 df, p=0.3
Score (logrank) test = 16.7 on 20 df, p=0.7, Robust = 20.1 p=0.4
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.90 1.48 1.79 1.92 2.26 5.41
[1] "Modality: WO residential\nTime interval: (30,50]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 374, number of events= 374
(332 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.266240 0.766255 0.333140 0.156403 -1.70
less_90d_tr1.lag -0.000868 0.999132 0.362284 0.230311 0.00
log_dias_treat_imp_sin_na.lag -0.074266 0.928425 0.072077 0.022583 -3.29
comp_bpsc_y3_severe.lag 0.172606 1.188397 0.384694 0.156994 1.10
policonsumo2.lag -0.110303 0.895562 0.413318 0.144697 -0.76
edad_al_ing_1 0.070663 1.073220 0.031269 0.032755 2.16
ano_nac_corr 0.070856 1.073426 0.030683 0.032237 2.20
susinidum_oh 1.203067 3.330315 0.559960 0.309989 3.88
susinidum_coc 1.489861 4.436480 0.605530 0.345834 4.31
susinidum_pbc 1.161622 3.195112 0.568769 0.322981 3.60
susinidum_mar 1.193610 3.298968 0.561149 0.308066 3.87
psycom_dum_with -0.067138 0.935066 0.150896 0.147822 -0.45
psycom_dum_study 0.000160 1.000160 0.172690 0.165963 0.00
freq_cons_dum_5day -0.025945 0.974389 0.124841 0.121203 -0.21
cond_oc_dum_2inact -0.166109 0.846954 0.226493 0.176276 -0.94
cond_oc_dum_3unemp -0.253565 0.776029 0.230443 0.191113 -1.33
susprindum_oh -0.862997 0.421896 0.396358 0.251589 -3.43
susprindum_coc -0.980906 0.374971 0.401764 0.255461 -3.84
susprindum_pbc -0.844257 0.429877 0.377333 0.219059 -3.85
susprindum_mar -0.321914 0.724761 0.572188 0.309357 -1.04
Pr(>|z|)
tr_outcome.lag 0.08871 .
less_90d_tr1.lag 0.99699
log_dias_treat_imp_sin_na.lag 0.00101 **
comp_bpsc_y3_severe.lag 0.27158
policonsumo2.lag 0.44588
edad_al_ing_1 0.03098 *
ano_nac_corr 0.02795 *
susinidum_oh 0.00010 ***
susinidum_coc 1.6e-05 ***
susinidum_pbc 0.00032 ***
susinidum_mar 0.00011 ***
psycom_dum_with 0.64970
psycom_dum_study 0.99923
freq_cons_dum_5day 0.83050
cond_oc_dum_2inact 0.34603
cond_oc_dum_3unemp 0.18458
susprindum_oh 0.00060 ***
susprindum_coc 0.00012 ***
susprindum_pbc 0.00012 ***
susprindum_mar 0.29807
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.766 1.305 0.564 1.041
less_90d_tr1.lag 0.999 1.001 0.636 1.569
log_dias_treat_imp_sin_na.lag 0.928 1.077 0.888 0.970
comp_bpsc_y3_severe.lag 1.188 0.841 0.874 1.617
policonsumo2.lag 0.896 1.117 0.674 1.189
edad_al_ing_1 1.073 0.932 1.006 1.144
ano_nac_corr 1.073 0.932 1.008 1.143
susinidum_oh 3.330 0.300 1.814 6.114
susinidum_coc 4.436 0.225 2.252 8.738
susinidum_pbc 3.195 0.313 1.697 6.017
susinidum_mar 3.299 0.303 1.804 6.034
psycom_dum_with 0.935 1.069 0.700 1.249
psycom_dum_study 1.000 1.000 0.722 1.385
freq_cons_dum_5day 0.974 1.026 0.768 1.236
cond_oc_dum_2inact 0.847 1.181 0.600 1.196
cond_oc_dum_3unemp 0.776 1.289 0.534 1.129
susprindum_oh 0.422 2.370 0.258 0.691
susprindum_coc 0.375 2.667 0.227 0.619
susprindum_pbc 0.430 2.326 0.280 0.660
susprindum_mar 0.725 1.380 0.395 1.329
Concordance= 0.559 (se = 0.016 )
Likelihood ratio test= 23.1 on 20 df, p=0.3
Wald test = 77 on 20 df, p=1e-08
Score (logrank) test = 22.5 on 20 df, p=0.3, Robust = 32.4 p=0.04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.28 0.80 0.91 0.95 1.07 3.45
[1] "Modality: GP intensive ambulatory\nTime interval: (30,50]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1813, number of events= 1813
(1667 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.06962 1.07210 0.17575 0.12489 0.56
less_90d_tr1.lag -0.07466 0.92806 0.20374 0.16249 -0.46
log_dias_treat_imp_sin_na.lag -0.09521 0.90918 0.04869 0.02817 -3.38
comp_bpsc_y3_severe.lag 0.13504 1.14459 0.16871 0.13019 1.04
policonsumo2.lag -0.12920 0.87880 0.18504 0.12763 -1.01
edad_al_ing_1 0.04836 1.04955 0.01269 0.01237 3.91
ano_nac_corr 0.05027 1.05156 0.01267 0.01256 4.00
susinidum_oh -0.22012 0.80242 0.18666 0.15267 -1.44
susinidum_coc -0.10184 0.90318 0.21646 0.18483 -0.55
susinidum_pbc -0.28630 0.75104 0.20610 0.18068 -1.58
susinidum_mar -0.26504 0.76718 0.18930 0.15669 -1.69
psycom_dum_with 0.03310 1.03365 0.05340 0.05364 0.62
psycom_dum_study -0.02673 0.97362 0.07087 0.07139 -0.37
freq_cons_dum_5day 0.02375 1.02404 0.04873 0.04830 0.49
cond_oc_dum_2inact -0.00351 0.99650 0.07187 0.07491 -0.05
cond_oc_dum_3unemp 0.09120 1.09549 0.05233 0.05132 1.78
susprindum_oh 0.03331 1.03387 0.21842 0.17404 0.19
susprindum_coc 0.08333 1.08691 0.22243 0.17940 0.46
susprindum_pbc 0.06182 1.06377 0.21718 0.17311 0.36
susprindum_mar 0.06484 1.06699 0.24196 0.19651 0.33
Pr(>|z|)
tr_outcome.lag 0.57722
less_90d_tr1.lag 0.64589
log_dias_treat_imp_sin_na.lag 0.00073 ***
comp_bpsc_y3_severe.lag 0.29961
policonsumo2.lag 0.31140
edad_al_ing_1 9.3e-05 ***
ano_nac_corr 6.2e-05 ***
susinidum_oh 0.14937
susinidum_coc 0.58165
susinidum_pbc 0.11306
susinidum_mar 0.09074 .
psycom_dum_with 0.53724
psycom_dum_study 0.70810
freq_cons_dum_5day 0.62291
cond_oc_dum_2inact 0.96262
cond_oc_dum_3unemp 0.07552 .
susprindum_oh 0.84823
susprindum_coc 0.64227
susprindum_pbc 0.72103
susprindum_mar 0.74142
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.072 0.933 0.839 1.369
less_90d_tr1.lag 0.928 1.078 0.675 1.276
log_dias_treat_imp_sin_na.lag 0.909 1.100 0.860 0.961
comp_bpsc_y3_severe.lag 1.145 0.874 0.887 1.477
policonsumo2.lag 0.879 1.138 0.684 1.129
edad_al_ing_1 1.050 0.953 1.024 1.075
ano_nac_corr 1.052 0.951 1.026 1.078
susinidum_oh 0.802 1.246 0.595 1.082
susinidum_coc 0.903 1.107 0.629 1.297
susinidum_pbc 0.751 1.331 0.527 1.070
susinidum_mar 0.767 1.303 0.564 1.043
psycom_dum_with 1.034 0.967 0.930 1.148
psycom_dum_study 0.974 1.027 0.846 1.120
freq_cons_dum_5day 1.024 0.977 0.932 1.126
cond_oc_dum_2inact 0.996 1.004 0.860 1.154
cond_oc_dum_3unemp 1.095 0.913 0.991 1.211
susprindum_oh 1.034 0.967 0.735 1.454
susprindum_coc 1.087 0.920 0.765 1.545
susprindum_pbc 1.064 0.940 0.758 1.493
susprindum_mar 1.067 0.937 0.726 1.568
Concordance= 0.541 (se = 0.008 )
Likelihood ratio test= 39.1 on 20 df, p=0.006
Wald test = 54.6 on 20 df, p=5e-05
Score (logrank) test = 38.1 on 20 df, p=0.009, Robust = 47.8 p=5e-04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.54 1.02 1.11 1.14 1.22 2.25
[1] "Modality: basic ambulatory\nTime interval: (30,50]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1568, number of events= 1568
(1465 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.2806 0.7553 0.2365 0.1623 -1.73
less_90d_tr1.lag -0.1925 0.8249 0.2784 0.2231 -0.86
log_dias_treat_imp_sin_na.lag -0.0676 0.9347 0.0771 0.0698 -0.97
comp_bpsc_y3_severe.lag 0.0964 1.1012 0.2221 0.1709 0.56
policonsumo2.lag 0.2001 1.2216 0.2863 0.2305 0.87
edad_al_ing_1 0.0626 1.0646 0.0137 0.0134 4.66
ano_nac_corr 0.0623 1.0643 0.0134 0.0134 4.63
susinidum_oh 0.2785 1.3211 0.1894 0.1996 1.40
susinidum_coc 0.1165 1.1236 0.2247 0.2413 0.48
susinidum_pbc 0.2363 1.2665 0.2212 0.2252 1.05
susinidum_mar 0.3165 1.3724 0.1923 0.2039 1.55
psycom_dum_with 0.0814 1.0848 0.0567 0.0577 1.41
psycom_dum_study 0.0763 1.0793 0.0745 0.0721 1.06
freq_cons_dum_5day -0.0662 0.9359 0.0585 0.0580 -1.14
cond_oc_dum_2inact -0.0688 0.9335 0.0799 0.0774 -0.89
cond_oc_dum_3unemp 0.0884 1.0924 0.0614 0.0612 1.44
susprindum_oh -0.1622 0.8503 0.2267 0.1865 -0.87
susprindum_coc -0.1144 0.8919 0.2313 0.1933 -0.59
susprindum_pbc -0.1726 0.8415 0.2259 0.1867 -0.92
susprindum_mar -0.2593 0.7716 0.2436 0.2006 -1.29
Pr(>|z|)
tr_outcome.lag 0.084 .
less_90d_tr1.lag 0.388
log_dias_treat_imp_sin_na.lag 0.333
comp_bpsc_y3_severe.lag 0.573
policonsumo2.lag 0.385
edad_al_ing_1 3.1e-06 ***
ano_nac_corr 3.6e-06 ***
susinidum_oh 0.163
susinidum_coc 0.629
susinidum_pbc 0.294
susinidum_mar 0.121
psycom_dum_with 0.159
psycom_dum_study 0.290
freq_cons_dum_5day 0.253
cond_oc_dum_2inact 0.374
cond_oc_dum_3unemp 0.149
susprindum_oh 0.385
susprindum_coc 0.554
susprindum_pbc 0.355
susprindum_mar 0.196
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.755 1.324 0.550 1.04
less_90d_tr1.lag 0.825 1.212 0.533 1.28
log_dias_treat_imp_sin_na.lag 0.935 1.070 0.815 1.07
comp_bpsc_y3_severe.lag 1.101 0.908 0.788 1.54
policonsumo2.lag 1.222 0.819 0.777 1.92
edad_al_ing_1 1.065 0.939 1.037 1.09
ano_nac_corr 1.064 0.940 1.037 1.09
susinidum_oh 1.321 0.757 0.893 1.95
susinidum_coc 1.124 0.890 0.700 1.80
susinidum_pbc 1.266 0.790 0.815 1.97
susinidum_mar 1.372 0.729 0.920 2.05
psycom_dum_with 1.085 0.922 0.969 1.21
psycom_dum_study 1.079 0.927 0.937 1.24
freq_cons_dum_5day 0.936 1.068 0.835 1.05
cond_oc_dum_2inact 0.933 1.071 0.802 1.09
cond_oc_dum_3unemp 1.092 0.915 0.969 1.23
susprindum_oh 0.850 1.176 0.590 1.23
susprindum_coc 0.892 1.121 0.611 1.30
susprindum_pbc 0.841 1.188 0.584 1.21
susprindum_mar 0.772 1.296 0.521 1.14
Concordance= 0.552 (se = 0.008 )
Likelihood ratio test= 46.9 on 20 df, p=6e-04
Wald test = 53.5 on 20 df, p=7e-05
Score (logrank) test = 45.8 on 20 df, p=9e-04, Robust = 50.2 p=2e-04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.564 0.755 0.854 0.876 0.966 1.782
[1] "Modality: GP residential\nTime interval: (30,50]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 756, number of events= 756
(698 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.11740 1.12457 0.27673 0.16989 0.69
less_90d_tr1.lag -0.26249 0.76914 0.35839 0.25620 -1.02
log_dias_treat_imp_sin_na.lag -0.03877 0.96197 0.09439 0.06274 -0.62
comp_bpsc_y3_severe.lag 0.83234 2.29869 0.35360 0.25940 3.21
policonsumo2.lag -0.60054 0.54852 0.33875 0.21316 -2.82
edad_al_ing_1 0.08079 1.08415 0.02030 0.01879 4.30
ano_nac_corr 0.08407 1.08771 0.02014 0.01864 4.51
susinidum_oh -0.02922 0.97120 0.24781 0.31946 -0.09
susinidum_coc 0.38760 1.47344 0.32554 0.36779 1.05
susinidum_pbc 0.10639 1.11225 0.26543 0.32825 0.32
susinidum_mar 0.00153 1.00153 0.24791 0.31782 0.00
psycom_dum_with -0.11299 0.89316 0.08404 0.08392 -1.35
psycom_dum_study -0.12624 0.88141 0.11726 0.11499 -1.10
freq_cons_dum_5day 0.00821 1.00824 0.08711 0.08190 0.10
cond_oc_dum_2inact 0.15730 1.17035 0.13868 0.14540 1.08
cond_oc_dum_3unemp 0.24127 1.27287 0.11430 0.10752 2.24
susprindum_oh 0.17025 1.18561 0.33711 0.28315 0.60
susprindum_coc -0.13689 0.87207 0.34443 0.29566 -0.46
susprindum_pbc -0.01076 0.98930 0.32630 0.27168 -0.04
susprindum_mar 0.12651 1.13486 0.37918 0.32173 0.39
Pr(>|z|)
tr_outcome.lag 0.4895
less_90d_tr1.lag 0.3056
log_dias_treat_imp_sin_na.lag 0.5366
comp_bpsc_y3_severe.lag 0.0013 **
policonsumo2.lag 0.0048 **
edad_al_ing_1 1.7e-05 ***
ano_nac_corr 6.5e-06 ***
susinidum_oh 0.9271
susinidum_coc 0.2919
susinidum_pbc 0.7459
susinidum_mar 0.9962
psycom_dum_with 0.1782
psycom_dum_study 0.2723
freq_cons_dum_5day 0.9201
cond_oc_dum_2inact 0.2793
cond_oc_dum_3unemp 0.0248 *
susprindum_oh 0.5476
susprindum_coc 0.6434
susprindum_pbc 0.9684
susprindum_mar 0.6942
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.125 0.889 0.806 1.569
less_90d_tr1.lag 0.769 1.300 0.466 1.271
log_dias_treat_imp_sin_na.lag 0.962 1.040 0.851 1.088
comp_bpsc_y3_severe.lag 2.299 0.435 1.383 3.822
policonsumo2.lag 0.549 1.823 0.361 0.833
edad_al_ing_1 1.084 0.922 1.045 1.125
ano_nac_corr 1.088 0.919 1.049 1.128
susinidum_oh 0.971 1.030 0.519 1.817
susinidum_coc 1.473 0.679 0.717 3.030
susinidum_pbc 1.112 0.899 0.585 2.116
susinidum_mar 1.002 0.998 0.537 1.867
psycom_dum_with 0.893 1.120 0.758 1.053
psycom_dum_study 0.881 1.135 0.704 1.104
freq_cons_dum_5day 1.008 0.992 0.859 1.184
cond_oc_dum_2inact 1.170 0.854 0.880 1.556
cond_oc_dum_3unemp 1.273 0.786 1.031 1.571
susprindum_oh 1.186 0.843 0.681 2.065
susprindum_coc 0.872 1.147 0.489 1.557
susprindum_pbc 0.989 1.011 0.581 1.685
susprindum_mar 1.135 0.881 0.604 2.132
Concordance= 0.568 (se = 0.012 )
Likelihood ratio test= 35.3 on 20 df, p=0.02
Wald test = 52.5 on 20 df, p=1e-04
Score (logrank) test = 35.3 on 20 df, p=0.02, Robust = 42.6 p=0.002
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.356 0.747 0.864 0.881 0.997 2.334
[1] "Modality: WO intensive ambulatory\nTime interval: (30,50]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 272, number of events= 272
(247 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.8814 0.4142 0.5441 0.5149 -1.71
less_90d_tr1.lag 1.5933 4.9200 0.6802 0.5705 2.79
log_dias_treat_imp_sin_na.lag 0.1368 1.1466 0.1408 0.0886 1.54
comp_bpsc_y3_severe.lag -0.8961 0.4082 0.5770 0.4669 -1.92
policonsumo2.lag 0.1904 1.2098 0.6121 0.5494 0.35
edad_al_ing_1 0.0287 1.0291 0.0334 0.0299 0.96
ano_nac_corr 0.0249 1.0252 0.0327 0.0303 0.82
susinidum_oh -0.5643 0.5688 0.4262 0.4298 -1.31
susinidum_coc -0.4115 0.6626 0.5134 0.5474 -0.75
susinidum_pbc -0.2852 0.7518 0.4498 0.4534 -0.63
susinidum_mar -0.3348 0.7155 0.4343 0.4430 -0.76
psycom_dum_with 0.1404 1.1507 0.1441 0.1395 1.01
psycom_dum_study -0.2148 0.8067 0.2120 0.1910 -1.12
freq_cons_dum_5day 0.0442 1.0452 0.1350 0.1374 0.32
cond_oc_dum_2inact 0.0642 1.0663 0.1672 0.1630 0.39
cond_oc_dum_3unemp 0.1041 1.1097 0.1727 0.1660 0.63
susprindum_oh -0.3823 0.6823 0.4451 0.2921 -1.31
susprindum_coc -0.1114 0.8946 0.4555 0.2828 -0.39
susprindum_pbc -0.3088 0.7343 0.4311 0.2677 -1.15
susprindum_mar -0.1062 0.8993 0.4811 0.3316 -0.32
Pr(>|z|)
tr_outcome.lag 0.0870 .
less_90d_tr1.lag 0.0052 **
log_dias_treat_imp_sin_na.lag 0.1225
comp_bpsc_y3_severe.lag 0.0550 .
policonsumo2.lag 0.7289
edad_al_ing_1 0.3377
ano_nac_corr 0.4116
susinidum_oh 0.1893
susinidum_coc 0.4522
susinidum_pbc 0.5293
susinidum_mar 0.4498
psycom_dum_with 0.3141
psycom_dum_study 0.2607
freq_cons_dum_5day 0.7477
cond_oc_dum_2inact 0.6937
cond_oc_dum_3unemp 0.5307
susprindum_oh 0.1906
susprindum_coc 0.6936
susprindum_pbc 0.2486
susprindum_mar 0.7489
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.414 2.414 0.151 1.14
less_90d_tr1.lag 4.920 0.203 1.608 15.05
log_dias_treat_imp_sin_na.lag 1.147 0.872 0.964 1.36
comp_bpsc_y3_severe.lag 0.408 2.450 0.163 1.02
policonsumo2.lag 1.210 0.827 0.412 3.55
edad_al_ing_1 1.029 0.972 0.971 1.09
ano_nac_corr 1.025 0.975 0.966 1.09
susinidum_oh 0.569 1.758 0.245 1.32
susinidum_coc 0.663 1.509 0.227 1.94
susinidum_pbc 0.752 1.330 0.309 1.83
susinidum_mar 0.715 1.398 0.300 1.70
psycom_dum_with 1.151 0.869 0.875 1.51
psycom_dum_study 0.807 1.240 0.555 1.17
freq_cons_dum_5day 1.045 0.957 0.798 1.37
cond_oc_dum_2inact 1.066 0.938 0.775 1.47
cond_oc_dum_3unemp 1.110 0.901 0.802 1.54
susprindum_oh 0.682 1.466 0.385 1.21
susprindum_coc 0.895 1.118 0.514 1.56
susprindum_pbc 0.734 1.362 0.435 1.24
susprindum_mar 0.899 1.112 0.469 1.72
Concordance= 0.565 (se = 0.018 )
Likelihood ratio test= 23.4 on 20 df, p=0.3
Wald test = 47.7 on 20 df, p=5e-04
Score (logrank) test = 21.3 on 20 df, p=0.4, Robust = 29 p=0.09
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.95 1.52 1.83 1.96 2.19 8.02
[1] "Modality: WO residential\nTime interval: (50,70]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 179, number of events= 179
(161 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.34247 0.71001 0.58467 0.53581 -0.64
less_90d_tr1.lag 0.21249 1.23676 0.75809 0.62317 0.34
log_dias_treat_imp_sin_na.lag -0.54691 0.57873 0.34815 0.24568 -2.23
comp_bpsc_y3_severe.lag 0.61124 1.84271 0.82117 0.69060 0.89
policonsumo2.lag 0.37707 1.45801 0.66872 0.38515 0.98
edad_al_ing_1 0.03296 1.03351 0.05659 0.05666 0.58
ano_nac_corr 0.04657 1.04767 0.05529 0.05527 0.84
susinidum_oh -0.86691 0.42025 0.63994 0.44744 -1.94
susinidum_coc -0.57378 0.56339 0.72954 0.51859 -1.11
susinidum_pbc -0.76562 0.46504 0.65980 0.46489 -1.65
susinidum_mar -0.41579 0.65982 0.63825 0.47410 -0.88
psycom_dum_with -0.57579 0.56226 0.22564 0.21500 -2.68
psycom_dum_study -0.64845 0.52285 0.26350 0.27055 -2.40
freq_cons_dum_5day 0.03824 1.03898 0.20734 0.20029 0.19
cond_oc_dum_2inact -0.10678 0.89872 0.30182 0.23316 -0.46
cond_oc_dum_3unemp -0.00802 0.99202 0.30611 0.23418 -0.03
susprindum_oh 0.02049 1.02070 0.52190 0.40816 0.05
susprindum_coc 0.24730 1.28057 0.53311 0.42121 0.59
susprindum_pbc 0.05301 1.05444 0.48619 0.38701 0.14
susprindum_mar -0.01869 0.98148 0.77774 0.59397 -0.03
Pr(>|z|)
tr_outcome.lag 0.5227
less_90d_tr1.lag 0.7331
log_dias_treat_imp_sin_na.lag 0.0260 *
comp_bpsc_y3_severe.lag 0.3761
policonsumo2.lag 0.3276
edad_al_ing_1 0.5607
ano_nac_corr 0.3995
susinidum_oh 0.0527 .
susinidum_coc 0.2685
susinidum_pbc 0.0996 .
susinidum_mar 0.3805
psycom_dum_with 0.0074 **
psycom_dum_study 0.0165 *
freq_cons_dum_5day 0.8486
cond_oc_dum_2inact 0.6470
cond_oc_dum_3unemp 0.9727
susprindum_oh 0.9600
susprindum_coc 0.5571
susprindum_pbc 0.8910
susprindum_mar 0.9749
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.710 1.408 0.248 2.029
less_90d_tr1.lag 1.237 0.809 0.365 4.195
log_dias_treat_imp_sin_na.lag 0.579 1.728 0.358 0.937
comp_bpsc_y3_severe.lag 1.843 0.543 0.476 7.133
policonsumo2.lag 1.458 0.686 0.685 3.102
edad_al_ing_1 1.034 0.968 0.925 1.155
ano_nac_corr 1.048 0.955 0.940 1.168
susinidum_oh 0.420 2.380 0.175 1.010
susinidum_coc 0.563 1.775 0.204 1.557
susinidum_pbc 0.465 2.150 0.187 1.157
susinidum_mar 0.660 1.516 0.261 1.671
psycom_dum_with 0.562 1.779 0.369 0.857
psycom_dum_study 0.523 1.913 0.308 0.889
freq_cons_dum_5day 1.039 0.962 0.702 1.538
cond_oc_dum_2inact 0.899 1.113 0.569 1.419
cond_oc_dum_3unemp 0.992 1.008 0.627 1.570
susprindum_oh 1.021 0.980 0.459 2.272
susprindum_coc 1.281 0.781 0.561 2.924
susprindum_pbc 1.054 0.948 0.494 2.251
susprindum_mar 0.981 1.019 0.306 3.144
Concordance= 0.606 (se = 0.025 )
Likelihood ratio test= 31.1 on 20 df, p=0.05
Wald test = 49.6 on 20 df, p=3e-04
Score (logrank) test = 29 on 20 df, p=0.09, Robust = 33.7 p=0.03
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.00 2.17 2.63 3.28 3.73 23.21
[1] "Modality: GP intensive ambulatory\nTime interval: (50,70]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 952, number of events= 952
(882 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.15073 1.16268 0.35885 0.20378 0.74
less_90d_tr1.lag -0.12136 0.88571 0.31729 0.20318 -0.60
log_dias_treat_imp_sin_na.lag -0.11381 0.89243 0.12990 0.08126 -1.40
comp_bpsc_y3_severe.lag 0.18725 1.20593 0.24138 0.15731 1.19
policonsumo2.lag -0.11770 0.88896 0.36364 0.21269 -0.55
edad_al_ing_1 0.09190 1.09625 0.02330 0.02268 4.05
ano_nac_corr 0.08750 1.09144 0.02299 0.02209 3.96
susinidum_oh -0.17827 0.83671 0.18565 0.18775 -0.95
susinidum_coc -0.00593 0.99409 0.23334 0.25203 -0.02
susinidum_pbc 0.09783 1.10277 0.22926 0.22434 0.44
susinidum_mar -0.05800 0.94365 0.19141 0.19611 -0.30
psycom_dum_with 0.11281 1.11942 0.07412 0.07576 1.49
psycom_dum_study 0.07157 1.07419 0.09999 0.09527 0.75
freq_cons_dum_5day 0.03258 1.03312 0.06785 0.06701 0.49
cond_oc_dum_2inact 0.08914 1.09324 0.10393 0.10346 0.86
cond_oc_dum_3unemp -0.00713 0.99289 0.07171 0.07102 -0.10
susprindum_oh -0.33803 0.71318 0.30827 0.26428 -1.28
susprindum_coc -0.49815 0.60766 0.31335 0.27108 -1.84
susprindum_pbc -0.35665 0.70002 0.30391 0.26104 -1.37
susprindum_mar -0.37735 0.68567 0.32531 0.29473 -1.28
Pr(>|z|)
tr_outcome.lag 0.460
less_90d_tr1.lag 0.550
log_dias_treat_imp_sin_na.lag 0.161
comp_bpsc_y3_severe.lag 0.234
policonsumo2.lag 0.580
edad_al_ing_1 5.1e-05 ***
ano_nac_corr 7.5e-05 ***
susinidum_oh 0.342
susinidum_coc 0.981
susinidum_pbc 0.663
susinidum_mar 0.767
psycom_dum_with 0.137
psycom_dum_study 0.453
freq_cons_dum_5day 0.627
cond_oc_dum_2inact 0.389
cond_oc_dum_3unemp 0.920
susprindum_oh 0.201
susprindum_coc 0.066 .
susprindum_pbc 0.172
susprindum_mar 0.200
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.163 0.860 0.780 1.73
less_90d_tr1.lag 0.886 1.129 0.595 1.32
log_dias_treat_imp_sin_na.lag 0.892 1.121 0.761 1.05
comp_bpsc_y3_severe.lag 1.206 0.829 0.886 1.64
policonsumo2.lag 0.889 1.125 0.586 1.35
edad_al_ing_1 1.096 0.912 1.049 1.15
ano_nac_corr 1.091 0.916 1.045 1.14
susinidum_oh 0.837 1.195 0.579 1.21
susinidum_coc 0.994 1.006 0.607 1.63
susinidum_pbc 1.103 0.907 0.710 1.71
susinidum_mar 0.944 1.060 0.643 1.39
psycom_dum_with 1.119 0.893 0.965 1.30
psycom_dum_study 1.074 0.931 0.891 1.29
freq_cons_dum_5day 1.033 0.968 0.906 1.18
cond_oc_dum_2inact 1.093 0.915 0.893 1.34
cond_oc_dum_3unemp 0.993 1.007 0.864 1.14
susprindum_oh 0.713 1.402 0.425 1.20
susprindum_coc 0.608 1.646 0.357 1.03
susprindum_pbc 0.700 1.429 0.420 1.17
susprindum_mar 0.686 1.458 0.385 1.22
Concordance= 0.563 (se = 0.01 )
Likelihood ratio test= 35.7 on 20 df, p=0.02
Wald test = 44.3 on 20 df, p=0.001
Score (logrank) test = 35.4 on 20 df, p=0.02, Robust = 41.4 p=0.003
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.81 1.28 1.46 1.50 1.68 3.37
[1] "Modality: basic ambulatory\nTime interval: (50,70]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 975, number of events= 975
(912 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.1636 0.8491 0.2844 0.2281 -0.72
less_90d_tr1.lag -0.4111 0.6629 0.3383 0.2087 -1.97
log_dias_treat_imp_sin_na.lag -0.1009 0.9041 0.0857 0.0621 -1.62
comp_bpsc_y3_severe.lag 0.2623 1.3000 0.2522 0.1943 1.35
policonsumo2.lag 0.0830 1.0866 0.2859 0.2368 0.35
edad_al_ing_1 0.1263 1.1346 0.0210 0.0197 6.41
ano_nac_corr 0.1315 1.1406 0.0209 0.0198 6.64
susinidum_oh -0.0636 0.9384 0.1742 0.2007 -0.32
susinidum_coc -0.3028 0.7387 0.2403 0.2725 -1.11
susinidum_pbc -0.1116 0.8944 0.2203 0.2294 -0.49
susinidum_mar -0.0607 0.9411 0.1806 0.2032 -0.30
psycom_dum_with -0.1183 0.8884 0.0739 0.0733 -1.61
psycom_dum_study -0.0289 0.9715 0.0939 0.0925 -0.31
freq_cons_dum_5day 0.0824 1.0859 0.0754 0.0759 1.09
cond_oc_dum_2inact 0.1289 1.1376 0.1038 0.1014 1.27
cond_oc_dum_3unemp 0.0419 1.0428 0.0784 0.0761 0.55
susprindum_oh 0.2948 1.3428 0.2325 0.2324 1.27
susprindum_coc 0.2484 1.2820 0.2364 0.2344 1.06
susprindum_pbc 0.1830 1.2009 0.2330 0.2318 0.79
susprindum_mar 0.0408 1.0416 0.2557 0.2599 0.16
Pr(>|z|)
tr_outcome.lag 0.473
less_90d_tr1.lag 0.049 *
log_dias_treat_imp_sin_na.lag 0.104
comp_bpsc_y3_severe.lag 0.177
policonsumo2.lag 0.726
edad_al_ing_1 1.4e-10 ***
ano_nac_corr 3.1e-11 ***
susinidum_oh 0.751
susinidum_coc 0.267
susinidum_pbc 0.627
susinidum_mar 0.765
psycom_dum_with 0.107
psycom_dum_study 0.754
freq_cons_dum_5day 0.278
cond_oc_dum_2inact 0.204
cond_oc_dum_3unemp 0.582
susprindum_oh 0.205
susprindum_coc 0.289
susprindum_pbc 0.430
susprindum_mar 0.875
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.849 1.178 0.543 1.328
less_90d_tr1.lag 0.663 1.508 0.440 0.998
log_dias_treat_imp_sin_na.lag 0.904 1.106 0.800 1.021
comp_bpsc_y3_severe.lag 1.300 0.769 0.888 1.903
policonsumo2.lag 1.087 0.920 0.683 1.728
edad_al_ing_1 1.135 0.881 1.092 1.179
ano_nac_corr 1.141 0.877 1.097 1.186
susinidum_oh 0.938 1.066 0.633 1.391
susinidum_coc 0.739 1.354 0.433 1.260
susinidum_pbc 0.894 1.118 0.570 1.402
susinidum_mar 0.941 1.063 0.632 1.401
psycom_dum_with 0.888 1.126 0.770 1.026
psycom_dum_study 0.971 1.029 0.810 1.164
freq_cons_dum_5day 1.086 0.921 0.936 1.260
cond_oc_dum_2inact 1.138 0.879 0.932 1.388
cond_oc_dum_3unemp 1.043 0.959 0.898 1.211
susprindum_oh 1.343 0.745 0.852 2.118
susprindum_coc 1.282 0.780 0.810 2.030
susprindum_pbc 1.201 0.833 0.762 1.891
susprindum_mar 1.042 0.960 0.626 1.734
Concordance= 0.578 (se = 0.01 )
Likelihood ratio test= 66 on 20 df, p=8e-07
Wald test = 70.7 on 20 df, p=1e-07
Score (logrank) test = 65.2 on 20 df, p=1e-06, Robust = 71 p=1e-07
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.468 0.703 0.866 0.903 1.059 2.559
[1] "Modality: GP residential\nTime interval: (50,70]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 392, number of events= 392
(361 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.9365 0.3920 0.4568 0.3123 -3.00
less_90d_tr1.lag 0.4865 1.6266 0.5871 0.4173 1.17
log_dias_treat_imp_sin_na.lag 0.0244 1.0247 0.1835 0.1265 0.19
comp_bpsc_y3_severe.lag -0.2099 0.8107 0.5023 0.3620 -0.58
policonsumo2.lag 0.2445 1.2769 0.6041 0.4989 0.49
edad_al_ing_1 0.0888 1.0928 0.0377 0.0376 2.36
ano_nac_corr 0.1002 1.1053 0.0376 0.0379 2.64
susinidum_oh 0.3006 1.3507 0.2867 0.2624 1.15
susinidum_coc 0.2230 1.2498 0.3793 0.3298 0.68
susinidum_pbc 0.1229 1.1308 0.3167 0.2983 0.41
susinidum_mar 0.0527 1.0541 0.2870 0.2583 0.20
psycom_dum_with -0.2135 0.8077 0.1207 0.1185 -1.80
psycom_dum_study -0.1890 0.8278 0.1750 0.1622 -1.17
freq_cons_dum_5day -0.1111 0.8948 0.1189 0.1213 -0.92
cond_oc_dum_2inact 0.0119 1.0120 0.2003 0.1955 0.06
cond_oc_dum_3unemp -0.1085 0.8972 0.1560 0.1579 -0.69
susprindum_oh -1.6833 0.1858 0.7330 0.2015 -8.35
susprindum_coc -2.0533 0.1283 0.7406 0.2254 -9.11
susprindum_pbc -1.8949 0.1503 0.7271 0.1580 -11.99
susprindum_mar -1.9872 0.1371 0.8088 0.4221 -4.71
Pr(>|z|)
tr_outcome.lag 0.0027 **
less_90d_tr1.lag 0.2437
log_dias_treat_imp_sin_na.lag 0.8471
comp_bpsc_y3_severe.lag 0.5621
policonsumo2.lag 0.6241
edad_al_ing_1 0.0183 *
ano_nac_corr 0.0083 **
susinidum_oh 0.2520
susinidum_coc 0.4989
susinidum_pbc 0.6803
susinidum_mar 0.8384
psycom_dum_with 0.0715 .
psycom_dum_study 0.2438
freq_cons_dum_5day 0.3595
cond_oc_dum_2inact 0.9514
cond_oc_dum_3unemp 0.4920
susprindum_oh < 2e-16 ***
susprindum_coc < 2e-16 ***
susprindum_pbc < 2e-16 ***
susprindum_mar 2.5e-06 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.392 2.551 0.2125 0.723
less_90d_tr1.lag 1.627 0.615 0.7179 3.685
log_dias_treat_imp_sin_na.lag 1.025 0.976 0.7997 1.313
comp_bpsc_y3_severe.lag 0.811 1.234 0.3987 1.648
policonsumo2.lag 1.277 0.783 0.4803 3.395
edad_al_ing_1 1.093 0.915 1.0152 1.176
ano_nac_corr 1.105 0.905 1.0261 1.191
susinidum_oh 1.351 0.740 0.8076 2.259
susinidum_coc 1.250 0.800 0.6548 2.386
susinidum_pbc 1.131 0.884 0.6301 2.029
susinidum_mar 1.054 0.949 0.6353 1.749
psycom_dum_with 0.808 1.238 0.6404 1.019
psycom_dum_study 0.828 1.208 0.6024 1.138
freq_cons_dum_5day 0.895 1.118 0.7055 1.135
cond_oc_dum_2inact 1.012 0.988 0.6899 1.484
cond_oc_dum_3unemp 0.897 1.115 0.6583 1.223
susprindum_oh 0.186 5.384 0.1252 0.276
susprindum_coc 0.128 7.794 0.0825 0.200
susprindum_pbc 0.150 6.652 0.1103 0.205
susprindum_mar 0.137 7.295 0.0599 0.314
Concordance= 0.578 (se = 0.017 )
Likelihood ratio test= 33.6 on 20 df, p=0.03
Wald test = 228 on 20 df, p=<2e-16
Score (logrank) test = 35 on 20 df, p=0.02, Robust = 32.5 p=0.04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.40 6.41 7.57 8.01 8.96 24.67
[1] "Modality: WO intensive ambulatory\nTime interval: (50,70]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 147, number of events= 147
(135 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 2.1717 8.7728 0.9544 0.7971 2.72
less_90d_tr1.lag 1.2465 3.4781 1.2201 0.7444 1.67
log_dias_treat_imp_sin_na.lag -0.2393 0.7872 0.1925 0.1025 -2.33
comp_bpsc_y3_severe.lag 2.0365 7.6639 1.0136 0.6358 3.20
policonsumo2.lag -2.9230 0.0538 1.2871 0.7707 -3.79
edad_al_ing_1 0.1148 1.1217 0.0580 0.0543 2.11
ano_nac_corr 0.1032 1.1087 0.0557 0.0526 1.96
susinidum_oh -0.3817 0.6827 0.6662 0.3756 -1.02
susinidum_coc 0.0769 1.0800 0.8669 0.7215 0.11
susinidum_pbc -0.6811 0.5060 0.7061 0.4589 -1.48
susinidum_mar -0.3223 0.7245 0.6791 0.4243 -0.76
psycom_dum_with -0.0156 0.9845 0.2213 0.2287 -0.07
psycom_dum_study 0.2414 1.2730 0.3154 0.3129 0.77
freq_cons_dum_5day -0.0729 0.9297 0.1940 0.1803 -0.40
cond_oc_dum_2inact 0.1242 1.1323 0.2438 0.2646 0.47
cond_oc_dum_3unemp -0.0392 0.9615 0.2570 0.2860 -0.14
susprindum_oh -0.5190 0.5951 0.7758 0.3672 -1.41
susprindum_coc -0.7877 0.4549 0.8090 0.5286 -1.49
susprindum_pbc -0.3925 0.6754 0.7660 0.3953 -0.99
susprindum_mar -1.0265 0.3583 0.8136 0.4512 -2.28
Pr(>|z|)
tr_outcome.lag 0.00644 **
less_90d_tr1.lag 0.09406 .
log_dias_treat_imp_sin_na.lag 0.01957 *
comp_bpsc_y3_severe.lag 0.00136 **
policonsumo2.lag 0.00015 ***
edad_al_ing_1 0.03457 *
ano_nac_corr 0.04983 *
susinidum_oh 0.30951
susinidum_coc 0.91510
susinidum_pbc 0.13773
susinidum_mar 0.44753
psycom_dum_with 0.94560
psycom_dum_study 0.44047
freq_cons_dum_5day 0.68583
cond_oc_dum_2inact 0.63867
cond_oc_dum_3unemp 0.89088
susprindum_oh 0.15756
susprindum_coc 0.13619
susprindum_pbc 0.32072
susprindum_mar 0.02290 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 8.7728 0.114 1.8393 41.844
less_90d_tr1.lag 3.4781 0.288 0.8085 14.963
log_dias_treat_imp_sin_na.lag 0.7872 1.270 0.6439 0.962
comp_bpsc_y3_severe.lag 7.6639 0.130 2.2042 26.647
policonsumo2.lag 0.0538 18.597 0.0119 0.244
edad_al_ing_1 1.1217 0.892 1.0084 1.248
ano_nac_corr 1.1087 0.902 1.0001 1.229
susinidum_oh 0.6827 1.465 0.3270 1.425
susinidum_coc 1.0800 0.926 0.2626 4.441
susinidum_pbc 0.5060 1.976 0.2059 1.244
susinidum_mar 0.7245 1.380 0.3154 1.664
psycom_dum_with 0.9845 1.016 0.6288 1.541
psycom_dum_study 1.2730 0.786 0.6894 2.351
freq_cons_dum_5day 0.9297 1.076 0.6529 1.324
cond_oc_dum_2inact 1.1323 0.883 0.6742 1.902
cond_oc_dum_3unemp 0.9615 1.040 0.5489 1.684
susprindum_oh 0.5951 1.680 0.2897 1.222
susprindum_coc 0.4549 2.198 0.1614 1.282
susprindum_pbc 0.6754 1.481 0.3112 1.466
susprindum_mar 0.3583 2.791 0.1480 0.867
Concordance= 0.616 (se = 0.024 )
Likelihood ratio test= 30.5 on 20 df, p=0.06
Wald test = 95.8 on 20 df, p=7e-12
Score (logrank) test = 28.2 on 20 df, p=0.1, Robust = 27.9 p=0.1
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.01 1.75 2.26 2.72 2.77 26.75
[1] "Modality: WO residential\nTime interval: (70,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 140, number of events= 140
(114 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.1218 1.1296 0.4994 0.3683 0.33
less_90d_tr1.lag -0.3431 0.7096 0.6058 0.3729 -0.92
log_dias_treat_imp_sin_na.lag -0.1413 0.8683 0.1374 0.1287 -1.10
comp_bpsc_y3_severe.lag -0.0120 0.9881 0.5646 0.3978 -0.03
policonsumo2.lag 0.3567 1.4286 0.6815 0.5466 0.65
edad_al_ing_1 0.7500 2.1170 0.1299 0.1256 5.97
ano_nac_corr 0.7538 2.1251 0.1269 0.1211 6.22
susinidum_oh -0.1588 0.8532 0.7420 0.6603 -0.24
susinidum_coc 0.2858 1.3308 0.9144 0.8027 0.36
susinidum_pbc 0.0273 1.0277 0.7797 0.7058 0.04
susinidum_mar -0.4422 0.6426 0.7490 0.6758 -0.65
psycom_dum_with 0.2162 1.2413 0.2481 0.2176 0.99
psycom_dum_study 0.3097 1.3631 0.3015 0.2705 1.15
freq_cons_dum_5day -0.1209 0.8861 0.2272 0.2189 -0.55
cond_oc_dum_2inact 0.3602 1.4335 0.3818 0.2755 1.31
cond_oc_dum_3unemp 0.3172 1.3732 0.3740 0.2661 1.19
susprindum_oh -0.8760 0.4165 0.6861 0.5430 -1.61
susprindum_coc -0.7450 0.4747 0.6871 0.5893 -1.26
susprindum_pbc -0.8698 0.4190 0.6718 0.5723 -1.52
susprindum_mar -0.0988 0.9060 0.8162 0.7142 -0.14
Pr(>|z|)
tr_outcome.lag 0.74
less_90d_tr1.lag 0.36
log_dias_treat_imp_sin_na.lag 0.27
comp_bpsc_y3_severe.lag 0.98
policonsumo2.lag 0.51
edad_al_ing_1 2.3e-09 ***
ano_nac_corr 4.9e-10 ***
susinidum_oh 0.81
susinidum_coc 0.72
susinidum_pbc 0.97
susinidum_mar 0.51
psycom_dum_with 0.32
psycom_dum_study 0.25
freq_cons_dum_5day 0.58
cond_oc_dum_2inact 0.19
cond_oc_dum_3unemp 0.23
susprindum_oh 0.11
susprindum_coc 0.21
susprindum_pbc 0.13
susprindum_mar 0.89
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.130 0.885 0.549 2.32
less_90d_tr1.lag 0.710 1.409 0.342 1.47
log_dias_treat_imp_sin_na.lag 0.868 1.152 0.675 1.12
comp_bpsc_y3_severe.lag 0.988 1.012 0.453 2.16
policonsumo2.lag 1.429 0.700 0.489 4.17
edad_al_ing_1 2.117 0.472 1.655 2.71
ano_nac_corr 2.125 0.471 1.676 2.69
susinidum_oh 0.853 1.172 0.234 3.11
susinidum_coc 1.331 0.751 0.276 6.42
susinidum_pbc 1.028 0.973 0.258 4.10
susinidum_mar 0.643 1.556 0.171 2.42
psycom_dum_with 1.241 0.806 0.810 1.90
psycom_dum_study 1.363 0.734 0.802 2.32
freq_cons_dum_5day 0.886 1.129 0.577 1.36
cond_oc_dum_2inact 1.434 0.698 0.835 2.46
cond_oc_dum_3unemp 1.373 0.728 0.815 2.31
susprindum_oh 0.416 2.401 0.144 1.21
susprindum_coc 0.475 2.106 0.150 1.51
susprindum_pbc 0.419 2.386 0.136 1.29
susprindum_mar 0.906 1.104 0.223 3.67
Concordance= 0.689 (se = 0.021 )
Likelihood ratio test= 59.3 on 20 df, p=9e-06
Wald test = 97 on 20 df, p=4e-12
Score (logrank) test = 66 on 20 df, p=8e-07, Robust = 66.5 p=7e-07
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.22 0.97 2.11 2.33 3.15 8.03
[1] "Modality: GP intensive ambulatory\nTime interval: (70,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 747, number of events= 747
(659 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.2388 0.7875 0.2250 0.1650 -1.45
less_90d_tr1.lag 0.0671 1.0694 0.3007 0.2596 0.26
log_dias_treat_imp_sin_na.lag -0.0766 0.9263 0.0842 0.0612 -1.25
comp_bpsc_y3_severe.lag 0.1116 1.1180 0.2199 0.1661 0.67
policonsumo2.lag 0.2374 1.2680 0.2585 0.1697 1.40
edad_al_ing_1 0.5374 1.7116 0.0388 0.0361 14.88
ano_nac_corr 0.5364 1.7098 0.0388 0.0359 14.95
susinidum_oh -0.0780 0.9250 0.2302 0.2226 -0.35
susinidum_coc 0.0180 1.0182 0.2923 0.2846 0.06
susinidum_pbc 0.0170 1.0171 0.2861 0.2684 0.06
susinidum_mar 0.0206 1.0208 0.2366 0.2294 0.09
psycom_dum_with -0.1143 0.8920 0.0833 0.0801 -1.43
psycom_dum_study -0.0224 0.9779 0.1145 0.1047 -0.21
freq_cons_dum_5day -0.0159 0.9842 0.0788 0.0728 -0.22
cond_oc_dum_2inact 0.0776 1.0807 0.1227 0.1112 0.70
cond_oc_dum_3unemp 0.0321 1.0326 0.0832 0.0774 0.41
susprindum_oh -0.0722 0.9304 0.3341 0.1880 -0.38
susprindum_coc -0.0727 0.9299 0.3364 0.1937 -0.38
susprindum_pbc -0.1282 0.8797 0.3291 0.1787 -0.72
susprindum_mar -0.0377 0.9630 0.3534 0.2304 -0.16
Pr(>|z|)
tr_outcome.lag 0.15
less_90d_tr1.lag 0.80
log_dias_treat_imp_sin_na.lag 0.21
comp_bpsc_y3_severe.lag 0.50
policonsumo2.lag 0.16
edad_al_ing_1 <2e-16 ***
ano_nac_corr <2e-16 ***
susinidum_oh 0.73
susinidum_coc 0.95
susinidum_pbc 0.95
susinidum_mar 0.93
psycom_dum_with 0.15
psycom_dum_study 0.83
freq_cons_dum_5day 0.83
cond_oc_dum_2inact 0.49
cond_oc_dum_3unemp 0.68
susprindum_oh 0.70
susprindum_coc 0.71
susprindum_pbc 0.47
susprindum_mar 0.87
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.788 1.270 0.570 1.09
less_90d_tr1.lag 1.069 0.935 0.643 1.78
log_dias_treat_imp_sin_na.lag 0.926 1.080 0.822 1.04
comp_bpsc_y3_severe.lag 1.118 0.894 0.807 1.55
policonsumo2.lag 1.268 0.789 0.909 1.77
edad_al_ing_1 1.712 0.584 1.595 1.84
ano_nac_corr 1.710 0.585 1.594 1.83
susinidum_oh 0.925 1.081 0.598 1.43
susinidum_coc 1.018 0.982 0.583 1.78
susinidum_pbc 1.017 0.983 0.601 1.72
susinidum_mar 1.021 0.980 0.651 1.60
psycom_dum_with 0.892 1.121 0.762 1.04
psycom_dum_study 0.978 1.023 0.796 1.20
freq_cons_dum_5day 0.984 1.016 0.853 1.14
cond_oc_dum_2inact 1.081 0.925 0.869 1.34
cond_oc_dum_3unemp 1.033 0.968 0.887 1.20
susprindum_oh 0.930 1.075 0.644 1.34
susprindum_coc 0.930 1.075 0.636 1.36
susprindum_pbc 0.880 1.137 0.620 1.25
susprindum_mar 0.963 1.038 0.613 1.51
Concordance= 0.642 (se = 0.011 )
Likelihood ratio test= 213 on 20 df, p=<2e-16
Wald test = 269 on 20 df, p=<2e-16
Score (logrank) test = 216 on 20 df, p=<2e-16, Robust = 195 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.19 0.77 1.29 1.46 1.93 6.42
[1] "Modality: basic ambulatory\nTime interval: (70,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 699, number of events= 699
(617 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.1225 1.1304 0.2986 0.2606 0.47
less_90d_tr1.lag -0.1555 0.8560 0.3452 0.2417 -0.64
log_dias_treat_imp_sin_na.lag -0.1626 0.8499 0.0931 0.0629 -2.58
comp_bpsc_y3_severe.lag 0.5189 1.6802 0.2279 0.2013 2.58
policonsumo2.lag 0.0739 1.0767 0.3118 0.2749 0.27
edad_al_ing_1 0.5115 1.6679 0.0417 0.0412 12.41
ano_nac_corr 0.5102 1.6655 0.0414 0.0410 12.44
susinidum_oh -0.1429 0.8668 0.2422 0.2256 -0.63
susinidum_coc -0.3566 0.7000 0.2801 0.2538 -1.40
susinidum_pbc -0.1280 0.8798 0.2786 0.2567 -0.50
susinidum_mar -0.1350 0.8737 0.2499 0.2303 -0.59
psycom_dum_with 0.0772 1.0803 0.0868 0.0828 0.93
psycom_dum_study -0.0475 0.9536 0.1097 0.0992 -0.48
freq_cons_dum_5day -0.0150 0.9851 0.0881 0.0823 -0.18
cond_oc_dum_2inact 0.1047 1.1104 0.1201 0.1228 0.85
cond_oc_dum_3unemp -0.1221 0.8850 0.0954 0.0850 -1.44
susprindum_oh -0.2467 0.7814 0.3190 0.2537 -0.97
susprindum_coc -0.1435 0.8663 0.3174 0.2495 -0.57
susprindum_pbc -0.1953 0.8226 0.3134 0.2427 -0.80
susprindum_mar -0.0947 0.9097 0.3344 0.2619 -0.36
Pr(>|z|)
tr_outcome.lag 0.6382
less_90d_tr1.lag 0.5201
log_dias_treat_imp_sin_na.lag 0.0098 **
comp_bpsc_y3_severe.lag 0.0100 **
policonsumo2.lag 0.7881
edad_al_ing_1 <2e-16 ***
ano_nac_corr <2e-16 ***
susinidum_oh 0.5265
susinidum_coc 0.1601
susinidum_pbc 0.6180
susinidum_mar 0.5576
psycom_dum_with 0.3511
psycom_dum_study 0.6320
freq_cons_dum_5day 0.8557
cond_oc_dum_2inact 0.3939
cond_oc_dum_3unemp 0.1507
susprindum_oh 0.3310
susprindum_coc 0.5653
susprindum_pbc 0.4210
susprindum_mar 0.7177
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.130 0.885 0.678 1.884
less_90d_tr1.lag 0.856 1.168 0.533 1.375
log_dias_treat_imp_sin_na.lag 0.850 1.177 0.751 0.962
comp_bpsc_y3_severe.lag 1.680 0.595 1.132 2.493
policonsumo2.lag 1.077 0.929 0.628 1.845
edad_al_ing_1 1.668 0.600 1.538 1.808
ano_nac_corr 1.666 0.600 1.537 1.805
susinidum_oh 0.867 1.154 0.557 1.349
susinidum_coc 0.700 1.428 0.426 1.151
susinidum_pbc 0.880 1.137 0.532 1.455
susinidum_mar 0.874 1.145 0.556 1.372
psycom_dum_with 1.080 0.926 0.918 1.271
psycom_dum_study 0.954 1.049 0.785 1.158
freq_cons_dum_5day 0.985 1.015 0.838 1.158
cond_oc_dum_2inact 1.110 0.901 0.873 1.413
cond_oc_dum_3unemp 0.885 1.130 0.749 1.045
susprindum_oh 0.781 1.280 0.475 1.285
susprindum_coc 0.866 1.154 0.531 1.413
susprindum_pbc 0.823 1.216 0.511 1.324
susprindum_mar 0.910 1.099 0.544 1.520
Concordance= 0.626 (se = 0.012 )
Likelihood ratio test= 168 on 20 df, p=<2e-16
Wald test = 186 on 20 df, p=<2e-16
Score (logrank) test = 168 on 20 df, p=<2e-16, Robust = 154 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.25 0.88 1.42 1.60 2.05 8.41
[1] "Modality: GP residential\nTime interval: (70,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 325, number of events= 325
(273 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.0928 0.9113 0.3333 0.2138 -0.43
less_90d_tr1.lag 0.0753 1.0782 0.3822 0.2800 0.27
log_dias_treat_imp_sin_na.lag -0.0724 0.9302 0.1271 0.1081 -0.67
comp_bpsc_y3_severe.lag 0.2079 1.2311 0.3464 0.2177 0.95
policonsumo2.lag -0.3087 0.7344 0.3645 0.2956 -1.04
edad_al_ing_1 0.6032 1.8279 0.0654 0.0575 10.49
ano_nac_corr 0.6012 1.8243 0.0655 0.0572 10.51
susinidum_oh -0.5745 0.5630 0.3975 0.2152 -2.67
susinidum_coc -0.5009 0.6060 0.4717 0.2745 -1.82
susinidum_pbc -0.6329 0.5310 0.4197 0.2457 -2.58
susinidum_mar -0.7080 0.4926 0.4014 0.2244 -3.16
psycom_dum_with 0.0821 1.0856 0.1289 0.1197 0.69
psycom_dum_study 0.1134 1.1201 0.1851 0.1634 0.69
freq_cons_dum_5day -0.1383 0.8709 0.1261 0.1162 -1.19
cond_oc_dum_2inact -0.3026 0.7389 0.2316 0.2019 -1.50
cond_oc_dum_3unemp -0.3872 0.6789 0.1644 0.1483 -2.61
susprindum_oh 0.1674 1.1822 0.7395 0.3219 0.52
susprindum_coc 0.2205 1.2467 0.7360 0.3168 0.70
susprindum_pbc 0.3005 1.3506 0.7227 0.2687 1.12
susprindum_mar 0.3069 1.3591 0.8012 0.3779 0.81
Pr(>|z|)
tr_outcome.lag 0.6641
less_90d_tr1.lag 0.7879
log_dias_treat_imp_sin_na.lag 0.5032
comp_bpsc_y3_severe.lag 0.3397
policonsumo2.lag 0.2964
edad_al_ing_1 <2e-16 ***
ano_nac_corr <2e-16 ***
susinidum_oh 0.0076 **
susinidum_coc 0.0681 .
susinidum_pbc 0.0100 *
susinidum_mar 0.0016 **
psycom_dum_with 0.4926
psycom_dum_study 0.4876
freq_cons_dum_5day 0.2340
cond_oc_dum_2inact 0.1339
cond_oc_dum_3unemp 0.0090 **
susprindum_oh 0.6031
susprindum_coc 0.4865
susprindum_pbc 0.2634
susprindum_mar 0.4167
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.911 1.097 0.599 1.386
less_90d_tr1.lag 1.078 0.927 0.623 1.867
log_dias_treat_imp_sin_na.lag 0.930 1.075 0.752 1.150
comp_bpsc_y3_severe.lag 1.231 0.812 0.803 1.886
policonsumo2.lag 0.734 1.362 0.411 1.311
edad_al_ing_1 1.828 0.547 1.633 2.046
ano_nac_corr 1.824 0.548 1.631 2.041
susinidum_oh 0.563 1.776 0.369 0.858
susinidum_coc 0.606 1.650 0.354 1.038
susinidum_pbc 0.531 1.883 0.328 0.860
susinidum_mar 0.493 2.030 0.317 0.765
psycom_dum_with 1.086 0.921 0.859 1.373
psycom_dum_study 1.120 0.893 0.813 1.543
freq_cons_dum_5day 0.871 1.148 0.694 1.094
cond_oc_dum_2inact 0.739 1.353 0.497 1.098
cond_oc_dum_3unemp 0.679 1.473 0.508 0.908
susprindum_oh 1.182 0.846 0.629 2.222
susprindum_coc 1.247 0.802 0.670 2.320
susprindum_pbc 1.351 0.740 0.798 2.287
susprindum_mar 1.359 0.736 0.648 2.850
Concordance= 0.651 (se = 0.015 )
Likelihood ratio test= 96.8 on 20 df, p=5e-12
Wald test = 144 on 20 df, p=<2e-16
Score (logrank) test = 100 on 20 df, p=1e-12, Robust = 103 p=4e-13
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.24 1.32 2.16 2.50 3.33 8.22
[1] "Modality: WO intensive ambulatory\nTime interval: (70,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 109, number of events= 109
(89 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.6955 2.0047 0.6932 0.5721 1.22
less_90d_tr1.lag 0.5006 1.6497 0.9757 0.8386 0.60
log_dias_treat_imp_sin_na.lag -0.0471 0.9540 0.1707 0.1345 -0.35
comp_bpsc_y3_severe.lag -0.4559 0.6339 0.6622 0.5470 -0.83
policonsumo2.lag -1.0245 0.3590 0.6121 0.4678 -2.19
edad_al_ing_1 0.7246 2.0639 0.1274 0.1228 5.90
ano_nac_corr 0.7012 2.0162 0.1237 0.1211 5.79
susinidum_oh 0.1961 1.2167 1.1315 0.4793 0.41
susinidum_coc -1.3926 0.2484 1.5539 0.5582 -2.49
susinidum_pbc 0.7068 2.0275 1.1711 0.5871 1.20
susinidum_mar 0.0357 1.0363 1.1510 0.5916 0.06
psycom_dum_with -0.0400 0.9608 0.2528 0.2598 -0.15
psycom_dum_study 0.3892 1.4758 0.3255 0.3183 1.22
freq_cons_dum_5day 0.4022 1.4951 0.2703 0.2447 1.64
cond_oc_dum_2inact 0.3064 1.3585 0.3531 0.3455 0.89
cond_oc_dum_3unemp 0.1412 1.1517 0.3483 0.3110 0.45
susprindum_oh 0.7067 2.0274 0.8058 0.6043 1.17
susprindum_coc 0.6402 1.8969 0.8214 0.6175 1.04
susprindum_pbc 0.6564 1.9279 0.7631 0.5779 1.14
susprindum_mar 1.1354 3.1123 0.8929 0.6353 1.79
Pr(>|z|)
tr_outcome.lag 0.224
less_90d_tr1.lag 0.551
log_dias_treat_imp_sin_na.lag 0.726
comp_bpsc_y3_severe.lag 0.405
policonsumo2.lag 0.029 *
edad_al_ing_1 3.6e-09 ***
ano_nac_corr 7.1e-09 ***
susinidum_oh 0.682
susinidum_coc 0.013 *
susinidum_pbc 0.229
susinidum_mar 0.952
psycom_dum_with 0.878
psycom_dum_study 0.221
freq_cons_dum_5day 0.100
cond_oc_dum_2inact 0.375
cond_oc_dum_3unemp 0.650
susprindum_oh 0.242
susprindum_coc 0.300
susprindum_pbc 0.256
susprindum_mar 0.074 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 2.005 0.499 0.6533 6.152
less_90d_tr1.lag 1.650 0.606 0.3188 8.535
log_dias_treat_imp_sin_na.lag 0.954 1.048 0.7329 1.242
comp_bpsc_y3_severe.lag 0.634 1.578 0.2170 1.852
policonsumo2.lag 0.359 2.786 0.1435 0.898
edad_al_ing_1 2.064 0.485 1.6223 2.626
ano_nac_corr 2.016 0.496 1.5901 2.556
susinidum_oh 1.217 0.822 0.4755 3.113
susinidum_coc 0.248 4.025 0.0832 0.742
susinidum_pbc 2.028 0.493 0.6416 6.408
susinidum_mar 1.036 0.965 0.3250 3.304
psycom_dum_with 0.961 1.041 0.5774 1.599
psycom_dum_study 1.476 0.678 0.7909 2.754
freq_cons_dum_5day 1.495 0.669 0.9255 2.415
cond_oc_dum_2inact 1.358 0.736 0.6901 2.674
cond_oc_dum_3unemp 1.152 0.868 0.6260 2.119
susprindum_oh 2.027 0.493 0.6202 6.627
susprindum_coc 1.897 0.527 0.5654 6.363
susprindum_pbc 1.928 0.519 0.6211 5.985
susprindum_mar 3.112 0.321 0.8961 10.810
Concordance= 0.691 (se = 0.024 )
Likelihood ratio test= 51.4 on 20 df, p=1e-04
Wald test = 286 on 20 df, p=<2e-16
Score (logrank) test = 49.1 on 20 df, p=3e-04, Robust = 44.4 p=0.001
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.01 0.16 0.30 0.55 0.58 9.39 Code
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:#:
#:#:#:#:#:#:#:#:#:#:#:
iiws_strat_alt_alt <- list()
comb_strata_alt_alt <- expand.grid(unique(data_mine_miss_restr_proc2$tipo_de_plan_2_mod),
unique(data_mine_miss_restr_proc2$time_interval3_alt))
for (ff in seq_along(1:nrow(comb_strata_alt_alt))) {
i<- comb_strata_alt_alt[ff,"Var1"]
j<- comb_strata_alt_alt[ff,"Var2"]
data_mine_miss_proc2_subset<-
subset(data_mine_miss_restr_proc2, subset= tipo_de_plan_2_mod==i & time_interval3_alt==j)
maxfu_df_subset <- subset(maxfu_restr_df, subset= maxfu.id %in% unique(data_mine_miss_proc2_subset$id))
iiw_model_strat_alt_alt<-
IrregLong::iiw.weights(Surv(time.lag,time,event)~
cluster(id)+ #If a frailty model is used, the cluster(id) term should appear before other covariates
tr_outcome.lag+
less_90d_tr1.lag+
log_dias_treat_imp_sin_na.lag +
comp_bpsc_y3_severe.lag+
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= data_mine_miss_proc2_subset,
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= maxfu_df_subset,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_oh", "susinidum_coc", "susinidum_pbc", "susinidum_mar", "psycom_dum_with", "psycom_dum_study_rec2", "freq_cons_dum_5day", "cond_oc_dum_3unemp", "cond_oc_dum_2inact", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "tr_outcome","log_dias_treat_imp_sin_na", "less_90d_tr1","comp_bpsc_y3_severe", "policonsumo2"),
lagfirst= c(2.95082,1,4.499811,1,1,1), #90/30.5 4.499811 es 90 días
first= T
)
iiws_strat_alt_alt[[ff]] <- list(
model = iiw_model_strat_alt_alt,
comb = paste0("Modality: ",i,"\nTime interval: ", j,"."),
cox_iiw = summary(iiw_model_strat_alt_alt$m),
iiw = summary(iiw_model_strat_alt_alt$iiw)
)
print(paste0("Modality: ",i,"\nTime interval: ", j))
print(summary(iiw_model_strat_alt_alt$m))
print(summary(iiw_model_strat_alt_alt$iiw))
}[1] "Modality: WO residential\nTime interval: [0,10]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 613, number of events= 613
(1197 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.0900 0.9140 0.2501 0.1810 -0.50
less_90d_tr1.lag 0.7064 2.0268 0.1946 0.1616 4.37
log_dias_treat_imp_sin_na.lag 0.0461 1.0472 0.0939 0.0916 0.50
comp_bpsc_y3_severe.lag 0.5370 1.7108 0.1989 0.1883 2.85
policonsumo2.lag -0.0548 0.9467 0.2159 0.2080 -0.26
edad_al_ing_1 -0.0295 0.9710 0.0169 0.0169 -1.74
ano_nac_corr -0.0248 0.9755 0.0162 0.0164 -1.51
susinidum_oh -0.0586 0.9430 0.2951 0.2920 -0.20
susinidum_coc -0.3375 0.7136 0.3417 0.3291 -1.03
susinidum_pbc 0.2985 1.3479 0.3018 0.3066 0.97
susinidum_mar 0.1756 1.1920 0.2911 0.2907 0.60
psycom_dum_with -0.2274 0.7966 0.1182 0.1159 -1.96
psycom_dum_study -0.2578 0.7728 0.1608 0.1667 -1.55
freq_cons_dum_5day 0.0151 1.0152 0.0988 0.0946 0.16
cond_oc_dum_2inact 0.2079 1.2311 0.1547 0.1475 1.41
cond_oc_dum_3unemp 0.1580 1.1712 0.1547 0.1502 1.05
susprindum_oh 0.3181 1.3746 0.2850 0.2514 1.27
susprindum_coc 0.4414 1.5549 0.2868 0.2377 1.86
susprindum_pbc 0.3105 1.3640 0.2662 0.2213 1.40
susprindum_mar 0.4231 1.5267 0.4242 0.3693 1.15
Pr(>|z|)
tr_outcome.lag 0.6192
less_90d_tr1.lag 1.2e-05 ***
log_dias_treat_imp_sin_na.lag 0.6147
comp_bpsc_y3_severe.lag 0.0044 **
policonsumo2.lag 0.7921
edad_al_ing_1 0.0815 .
ano_nac_corr 0.1308
susinidum_oh 0.8409
susinidum_coc 0.3052
susinidum_pbc 0.3302
susinidum_mar 0.5457
psycom_dum_with 0.0497 *
psycom_dum_study 0.1219
freq_cons_dum_5day 0.8730
cond_oc_dum_2inact 0.1588
cond_oc_dum_3unemp 0.2928
susprindum_oh 0.2057
susprindum_coc 0.0633 .
susprindum_pbc 0.1607
susprindum_mar 0.2519
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.914 1.094 0.641 1.30
less_90d_tr1.lag 2.027 0.493 1.477 2.78
log_dias_treat_imp_sin_na.lag 1.047 0.955 0.875 1.25
comp_bpsc_y3_severe.lag 1.711 0.585 1.183 2.47
policonsumo2.lag 0.947 1.056 0.630 1.42
edad_al_ing_1 0.971 1.030 0.939 1.00
ano_nac_corr 0.976 1.025 0.945 1.01
susinidum_oh 0.943 1.060 0.532 1.67
susinidum_coc 0.714 1.401 0.374 1.36
susinidum_pbc 1.348 0.742 0.739 2.46
susinidum_mar 1.192 0.839 0.674 2.11
psycom_dum_with 0.797 1.255 0.635 1.00
psycom_dum_study 0.773 1.294 0.557 1.07
freq_cons_dum_5day 1.015 0.985 0.843 1.22
cond_oc_dum_2inact 1.231 0.812 0.922 1.64
cond_oc_dum_3unemp 1.171 0.854 0.873 1.57
susprindum_oh 1.375 0.727 0.840 2.25
susprindum_coc 1.555 0.643 0.976 2.48
susprindum_pbc 1.364 0.733 0.884 2.10
susprindum_mar 1.527 0.655 0.740 3.15
Concordance= 0.603 (se = 0.012 )
Likelihood ratio test= 62.1 on 20 df, p=3e-06
Wald test = 79.2 on 20 df, p=5e-09
Score (logrank) test = 57.3 on 20 df, p=2e-05, Robust = 63.2 p=2e-06
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.125 0.197 0.226 0.251 0.280 1.222
[1] "Modality: GP intensive ambulatory\nTime interval: [0,10]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 3052, number of events= 3052
(5099 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.04944 0.95177 0.19147 0.14628 -0.34
less_90d_tr1.lag 0.37121 1.44949 0.09331 0.08001 4.64
log_dias_treat_imp_sin_na.lag 0.12986 1.13867 0.05504 0.05510 2.36
comp_bpsc_y3_severe.lag 0.58095 1.78773 0.07941 0.07147 8.13
policonsumo2.lag 0.09846 1.10347 0.12622 0.09638 1.02
edad_al_ing_1 0.01296 1.01304 0.00803 0.00836 1.55
ano_nac_corr 0.01642 1.01655 0.00788 0.00829 1.98
susinidum_oh 0.04711 1.04824 0.13295 0.12352 0.38
susinidum_coc 0.08476 1.08845 0.15803 0.14920 0.57
susinidum_pbc 0.24050 1.27188 0.14782 0.14249 1.69
susinidum_mar 0.09664 1.10146 0.13417 0.12551 0.77
psycom_dum_with -0.06549 0.93661 0.03946 0.03958 -1.65
psycom_dum_study 0.30426 1.35562 0.06520 0.07476 4.07
freq_cons_dum_5day 0.03776 1.03848 0.03760 0.03824 0.99
cond_oc_dum_2inact -0.00650 0.99353 0.05709 0.05742 -0.11
cond_oc_dum_3unemp 0.00356 1.00356 0.04041 0.04027 0.09
susprindum_oh 0.08893 1.09301 0.17883 0.17307 0.51
susprindum_coc 0.05586 1.05745 0.17922 0.17480 0.32
susprindum_pbc 0.13574 1.14539 0.17488 0.17189 0.79
susprindum_mar 0.07097 1.07355 0.18876 0.18724 0.38
Pr(>|z|)
tr_outcome.lag 0.735
less_90d_tr1.lag 3.5e-06 ***
log_dias_treat_imp_sin_na.lag 0.018 *
comp_bpsc_y3_severe.lag 4.4e-16 ***
policonsumo2.lag 0.307
edad_al_ing_1 0.121
ano_nac_corr 0.048 *
susinidum_oh 0.703
susinidum_coc 0.570
susinidum_pbc 0.091 .
susinidum_mar 0.441
psycom_dum_with 0.098 .
psycom_dum_study 4.7e-05 ***
freq_cons_dum_5day 0.323
cond_oc_dum_2inact 0.910
cond_oc_dum_3unemp 0.930
susprindum_oh 0.607
susprindum_coc 0.749
susprindum_pbc 0.430
susprindum_mar 0.705
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.952 1.051 0.715 1.27
less_90d_tr1.lag 1.449 0.690 1.239 1.70
log_dias_treat_imp_sin_na.lag 1.139 0.878 1.022 1.27
comp_bpsc_y3_severe.lag 1.788 0.559 1.554 2.06
policonsumo2.lag 1.103 0.906 0.914 1.33
edad_al_ing_1 1.013 0.987 0.997 1.03
ano_nac_corr 1.017 0.984 1.000 1.03
susinidum_oh 1.048 0.954 0.823 1.34
susinidum_coc 1.088 0.919 0.812 1.46
susinidum_pbc 1.272 0.786 0.962 1.68
susinidum_mar 1.101 0.908 0.861 1.41
psycom_dum_with 0.937 1.068 0.867 1.01
psycom_dum_study 1.356 0.738 1.171 1.57
freq_cons_dum_5day 1.038 0.963 0.963 1.12
cond_oc_dum_2inact 0.994 1.007 0.888 1.11
cond_oc_dum_3unemp 1.004 0.996 0.927 1.09
susprindum_oh 1.093 0.915 0.779 1.53
susprindum_coc 1.057 0.946 0.751 1.49
susprindum_pbc 1.145 0.873 0.818 1.60
susprindum_mar 1.074 0.931 0.744 1.55
Concordance= 0.585 (se = 0.006 )
Likelihood ratio test= 156 on 20 df, p=<2e-16
Wald test = 154 on 20 df, p=<2e-16
Score (logrank) test = 141 on 20 df, p=<2e-16, Robust = 139 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.164 0.255 0.304 0.308 0.331 1.985
[1] "Modality: basic ambulatory\nTime interval: [0,10]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 2876, number of events= 2876
(4323 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.05522 1.05677 0.25750 0.21420 0.26
less_90d_tr1.lag 0.30335 1.35439 0.11227 0.10746 2.82
log_dias_treat_imp_sin_na.lag 0.01684 1.01699 0.06516 0.05205 0.32
comp_bpsc_y3_severe.lag 0.70747 2.02885 0.08855 0.08420 8.40
policonsumo2.lag -0.01761 0.98254 0.12635 0.10821 -0.16
edad_al_ing_1 0.04075 1.04159 0.00875 0.00923 4.42
ano_nac_corr 0.04753 1.04867 0.00853 0.00912 5.21
susinidum_oh 0.17495 1.19119 0.14627 0.19430 0.90
susinidum_coc 0.38052 1.46304 0.16633 0.21181 1.80
susinidum_pbc 0.22510 1.25244 0.16317 0.20567 1.09
susinidum_mar 0.26295 1.30076 0.14697 0.19576 1.34
psycom_dum_with -0.05866 0.94303 0.04148 0.04207 -1.39
psycom_dum_study 0.46136 1.58624 0.05948 0.06699 6.89
freq_cons_dum_5day 0.05859 1.06034 0.04306 0.04442 1.32
cond_oc_dum_2inact 0.09190 1.09626 0.06201 0.06085 1.51
cond_oc_dum_3unemp 0.03188 1.03239 0.04548 0.04753 0.67
susprindum_oh 0.48317 1.62121 0.22515 0.20340 2.38
susprindum_coc 0.56258 1.75519 0.22597 0.20514 2.74
susprindum_pbc 0.48472 1.62372 0.22336 0.20300 2.39
susprindum_mar 0.44114 1.55448 0.23236 0.20939 2.11
Pr(>|z|)
tr_outcome.lag 0.7966
less_90d_tr1.lag 0.0048 **
log_dias_treat_imp_sin_na.lag 0.7462
comp_bpsc_y3_severe.lag < 2e-16 ***
policonsumo2.lag 0.8707
edad_al_ing_1 1.0e-05 ***
ano_nac_corr 1.8e-07 ***
susinidum_oh 0.3679
susinidum_coc 0.0724 .
susinidum_pbc 0.2738
susinidum_mar 0.1792
psycom_dum_with 0.1632
psycom_dum_study 5.7e-12 ***
freq_cons_dum_5day 0.1871
cond_oc_dum_2inact 0.1310
cond_oc_dum_3unemp 0.5024
susprindum_oh 0.0175 *
susprindum_coc 0.0061 **
susprindum_pbc 0.0169 *
susprindum_mar 0.0351 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.057 0.946 0.694 1.61
less_90d_tr1.lag 1.354 0.738 1.097 1.67
log_dias_treat_imp_sin_na.lag 1.017 0.983 0.918 1.13
comp_bpsc_y3_severe.lag 2.029 0.493 1.720 2.39
policonsumo2.lag 0.983 1.018 0.795 1.21
edad_al_ing_1 1.042 0.960 1.023 1.06
ano_nac_corr 1.049 0.954 1.030 1.07
susinidum_oh 1.191 0.839 0.814 1.74
susinidum_coc 1.463 0.684 0.966 2.22
susinidum_pbc 1.252 0.798 0.837 1.87
susinidum_mar 1.301 0.769 0.886 1.91
psycom_dum_with 0.943 1.060 0.868 1.02
psycom_dum_study 1.586 0.630 1.391 1.81
freq_cons_dum_5day 1.060 0.943 0.972 1.16
cond_oc_dum_2inact 1.096 0.912 0.973 1.24
cond_oc_dum_3unemp 1.032 0.969 0.941 1.13
susprindum_oh 1.621 0.617 1.088 2.42
susprindum_coc 1.755 0.570 1.174 2.62
susprindum_pbc 1.624 0.616 1.091 2.42
susprindum_mar 1.554 0.643 1.031 2.34
Concordance= 0.598 (se = 0.006 )
Likelihood ratio test= 222 on 20 df, p=<2e-16
Wald test = 212 on 20 df, p=<2e-16
Score (logrank) test = 208 on 20 df, p=<2e-16, Robust = 200 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.060 0.132 0.165 0.170 0.188 1.000
[1] "Modality: GP residential\nTime interval: [0,10]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1378, number of events= 1378
(2346 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.15244 1.16468 0.22389 0.19303 0.79
less_90d_tr1.lag 0.33669 1.40030 0.12075 0.09077 3.71
log_dias_treat_imp_sin_na.lag 0.05576 1.05735 0.05762 0.06101 0.91
comp_bpsc_y3_severe.lag 0.23214 1.26129 0.14637 0.09912 2.34
policonsumo2.lag 0.00257 1.00258 0.17348 0.14029 0.02
edad_al_ing_1 0.01254 1.01262 0.01203 0.01134 1.11
ano_nac_corr 0.01722 1.01737 0.01179 0.01141 1.51
susinidum_oh 0.32715 1.38701 0.21824 0.18145 1.80
susinidum_coc 0.31906 1.37584 0.25651 0.22809 1.40
susinidum_pbc 0.28383 1.32820 0.23008 0.19598 1.45
susinidum_mar 0.36668 1.44294 0.21914 0.18275 2.01
psycom_dum_with -0.08436 0.91910 0.06180 0.06083 -1.39
psycom_dum_study 0.05554 1.05712 0.09488 0.10204 0.54
freq_cons_dum_5day 0.14407 1.15497 0.06361 0.06306 2.28
cond_oc_dum_2inact 0.00969 1.00974 0.10245 0.10074 0.10
cond_oc_dum_3unemp -0.01209 0.98799 0.08698 0.08737 -0.14
susprindum_oh 0.22130 1.24769 0.29117 0.17490 1.27
susprindum_coc 0.12158 1.12928 0.29393 0.18007 0.68
susprindum_pbc 0.25572 1.29139 0.28565 0.16995 1.50
susprindum_mar 0.10095 1.10622 0.33636 0.23222 0.43
Pr(>|z|)
tr_outcome.lag 0.42968
less_90d_tr1.lag 0.00021 ***
log_dias_treat_imp_sin_na.lag 0.36073
comp_bpsc_y3_severe.lag 0.01919 *
policonsumo2.lag 0.98537
edad_al_ing_1 0.26872
ano_nac_corr 0.13110
susinidum_oh 0.07139 .
susinidum_coc 0.16185
susinidum_pbc 0.14755
susinidum_mar 0.04480 *
psycom_dum_with 0.16553
psycom_dum_study 0.58621
freq_cons_dum_5day 0.02232 *
cond_oc_dum_2inact 0.92334
cond_oc_dum_3unemp 0.88998
susprindum_oh 0.20577
susprindum_coc 0.49955
susprindum_pbc 0.13240
susprindum_mar 0.66377
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.165 0.859 0.798 1.70
less_90d_tr1.lag 1.400 0.714 1.172 1.67
log_dias_treat_imp_sin_na.lag 1.057 0.946 0.938 1.19
comp_bpsc_y3_severe.lag 1.261 0.793 1.039 1.53
policonsumo2.lag 1.003 0.997 0.762 1.32
edad_al_ing_1 1.013 0.988 0.990 1.04
ano_nac_corr 1.017 0.983 0.995 1.04
susinidum_oh 1.387 0.721 0.972 1.98
susinidum_coc 1.376 0.727 0.880 2.15
susinidum_pbc 1.328 0.753 0.905 1.95
susinidum_mar 1.443 0.693 1.009 2.06
psycom_dum_with 0.919 1.088 0.816 1.04
psycom_dum_study 1.057 0.946 0.865 1.29
freq_cons_dum_5day 1.155 0.866 1.021 1.31
cond_oc_dum_2inact 1.010 0.990 0.829 1.23
cond_oc_dum_3unemp 0.988 1.012 0.832 1.17
susprindum_oh 1.248 0.801 0.886 1.76
susprindum_coc 1.129 0.886 0.793 1.61
susprindum_pbc 1.291 0.774 0.926 1.80
susprindum_mar 1.106 0.904 0.702 1.74
Concordance= 0.567 (se = 0.009 )
Likelihood ratio test= 40.4 on 20 df, p=0.004
Wald test = 49.6 on 20 df, p=3e-04
Score (logrank) test = 38.5 on 20 df, p=0.008, Robust = 42.4 p=0.002
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.184 0.233 0.256 0.265 0.284 0.584
[1] "Modality: WO intensive ambulatory\nTime interval: [0,10]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 439, number of events= 439
(710 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.7998 0.4494 1.0298 0.2328 -3.44
less_90d_tr1.lag 0.8186 2.2672 0.2904 0.2384 3.43
log_dias_treat_imp_sin_na.lag 0.3638 1.4389 0.2272 0.1822 2.00
comp_bpsc_y3_severe.lag 0.1953 1.2157 0.2301 0.1749 1.12
policonsumo2.lag 0.6427 1.9016 0.3232 0.2527 2.54
edad_al_ing_1 0.0258 1.0261 0.0237 0.0230 1.12
ano_nac_corr 0.0237 1.0239 0.0235 0.0231 1.02
susinidum_oh -0.0563 0.9453 0.3381 0.4084 -0.14
susinidum_coc -0.1008 0.9042 0.3994 0.4504 -0.22
susinidum_pbc -0.2883 0.7496 0.3687 0.4356 -0.66
susinidum_mar -0.1510 0.8598 0.3476 0.4131 -0.37
psycom_dum_with 0.2137 1.2383 0.1094 0.1041 2.05
psycom_dum_study 0.6845 1.9827 0.1883 0.1839 3.72
freq_cons_dum_5day 0.0397 1.0405 0.1021 0.1033 0.38
cond_oc_dum_2inact 0.1214 1.1291 0.1298 0.1209 1.00
cond_oc_dum_3unemp 0.1248 1.1329 0.1386 0.1308 0.95
susprindum_oh 0.3577 1.4300 0.3910 0.2730 1.31
susprindum_coc 0.6496 1.9149 0.3979 0.2808 2.31
susprindum_pbc 0.6304 1.8784 0.3826 0.2603 2.42
susprindum_mar 1.1684 3.2168 0.4236 0.2973 3.93
Pr(>|z|)
tr_outcome.lag 0.00059 ***
less_90d_tr1.lag 0.00060 ***
log_dias_treat_imp_sin_na.lag 0.04584 *
comp_bpsc_y3_severe.lag 0.26411
policonsumo2.lag 0.01098 *
edad_al_ing_1 0.26148
ano_nac_corr 0.30580
susinidum_oh 0.89044
susinidum_coc 0.82298
susinidum_pbc 0.50811
susinidum_mar 0.71473
psycom_dum_with 0.04000 *
psycom_dum_study 0.00020 ***
freq_cons_dum_5day 0.70098
cond_oc_dum_2inact 0.31526
cond_oc_dum_3unemp 0.34005
susprindum_oh 0.19007
susprindum_coc 0.02069 *
susprindum_pbc 0.01542 *
susprindum_mar 8.5e-05 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.449 2.225 0.285 0.709
less_90d_tr1.lag 2.267 0.441 1.421 3.617
log_dias_treat_imp_sin_na.lag 1.439 0.695 1.007 2.056
comp_bpsc_y3_severe.lag 1.216 0.823 0.863 1.713
policonsumo2.lag 1.902 0.526 1.159 3.121
edad_al_ing_1 1.026 0.975 0.981 1.073
ano_nac_corr 1.024 0.977 0.979 1.071
susinidum_oh 0.945 1.058 0.425 2.105
susinidum_coc 0.904 1.106 0.374 2.186
susinidum_pbc 0.750 1.334 0.319 1.760
susinidum_mar 0.860 1.163 0.383 1.932
psycom_dum_with 1.238 0.808 1.010 1.519
psycom_dum_study 1.983 0.504 1.383 2.843
freq_cons_dum_5day 1.040 0.961 0.850 1.274
cond_oc_dum_2inact 1.129 0.886 0.891 1.431
cond_oc_dum_3unemp 1.133 0.883 0.877 1.464
susprindum_oh 1.430 0.699 0.838 2.442
susprindum_coc 1.915 0.522 1.104 3.320
susprindum_pbc 1.878 0.532 1.128 3.128
susprindum_mar 3.217 0.311 1.796 5.761
Concordance= 0.612 (se = 0.015 )
Likelihood ratio test= 42.4 on 20 df, p=0.002
Wald test = 83.1 on 20 df, p=1e-09
Score (logrank) test = 39.6 on 20 df, p=0.006, Robust = 52.3 p=1e-04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.059 0.145 0.190 0.201 0.230 1.315
[1] "Modality: WO residential\nTime interval: (10,20]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 510, number of events= 510
(460 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.21592 1.24101 0.34608 0.31001 0.70
less_90d_tr1.lag 0.89083 2.43714 0.41911 0.35694 2.50
log_dias_treat_imp_sin_na.lag 0.34591 1.41328 0.31806 0.22707 1.52
comp_bpsc_y3_severe.lag -0.11404 0.89222 0.34659 0.28279 -0.40
policonsumo2.lag -0.09157 0.91250 0.45600 0.21426 -0.43
edad_al_ing_1 -0.01132 0.98874 0.01971 0.01840 -0.62
ano_nac_corr -0.01574 0.98439 0.01894 0.01751 -0.90
susinidum_oh -0.30205 0.73930 0.29743 0.42830 -0.71
susinidum_coc -0.26431 0.76773 0.34479 0.45702 -0.58
susinidum_pbc -0.51114 0.59981 0.30812 0.44117 -1.16
susinidum_mar -0.31219 0.73184 0.29812 0.42846 -0.73
psycom_dum_with 0.09827 1.10326 0.13967 0.13716 0.72
psycom_dum_study 0.06058 1.06245 0.16328 0.16532 0.37
freq_cons_dum_5day 0.04197 1.04286 0.10563 0.10603 0.40
cond_oc_dum_2inact 0.09249 1.09690 0.16558 0.15013 0.62
cond_oc_dum_3unemp -0.04958 0.95162 0.16511 0.15264 -0.32
susprindum_oh 0.13041 1.13929 0.29484 0.29199 0.45
susprindum_coc -0.31573 0.72926 0.30716 0.31845 -0.99
susprindum_pbc 0.00569 1.00571 0.28545 0.28959 0.02
susprindum_mar -0.87667 0.41617 0.53610 0.39618 -2.21
Pr(>|z|)
tr_outcome.lag 0.486
less_90d_tr1.lag 0.013 *
log_dias_treat_imp_sin_na.lag 0.128
comp_bpsc_y3_severe.lag 0.687
policonsumo2.lag 0.669
edad_al_ing_1 0.538
ano_nac_corr 0.369
susinidum_oh 0.481
susinidum_coc 0.563
susinidum_pbc 0.247
susinidum_mar 0.466
psycom_dum_with 0.474
psycom_dum_study 0.714
freq_cons_dum_5day 0.692
cond_oc_dum_2inact 0.538
cond_oc_dum_3unemp 0.745
susprindum_oh 0.655
susprindum_coc 0.321
susprindum_pbc 0.984
susprindum_mar 0.027 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.241 0.806 0.676 2.279
less_90d_tr1.lag 2.437 0.410 1.211 4.906
log_dias_treat_imp_sin_na.lag 1.413 0.708 0.906 2.205
comp_bpsc_y3_severe.lag 0.892 1.121 0.513 1.553
policonsumo2.lag 0.913 1.096 0.600 1.389
edad_al_ing_1 0.989 1.011 0.954 1.025
ano_nac_corr 0.984 1.016 0.951 1.019
susinidum_oh 0.739 1.353 0.319 1.712
susinidum_coc 0.768 1.303 0.313 1.880
susinidum_pbc 0.600 1.667 0.253 1.424
susinidum_mar 0.732 1.366 0.316 1.695
psycom_dum_with 1.103 0.906 0.843 1.444
psycom_dum_study 1.062 0.941 0.768 1.469
freq_cons_dum_5day 1.043 0.959 0.847 1.284
cond_oc_dum_2inact 1.097 0.912 0.817 1.472
cond_oc_dum_3unemp 0.952 1.051 0.706 1.283
susprindum_oh 1.139 0.878 0.643 2.019
susprindum_coc 0.729 1.371 0.391 1.361
susprindum_pbc 1.006 0.994 0.570 1.774
susprindum_mar 0.416 2.403 0.191 0.905
Concordance= 0.563 (se = 0.014 )
Likelihood ratio test= 31.6 on 20 df, p=0.05
Wald test = 48.2 on 20 df, p=4e-04
Score (logrank) test = 29.6 on 20 df, p=0.08, Robust = 36.9 p=0.01
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.318 0.466 0.529 0.571 0.630 1.853
[1] "Modality: GP intensive ambulatory\nTime interval: (10,20]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 2228, number of events= 2228
(2101 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.21836 0.80384 0.20606 0.15412 -1.42
less_90d_tr1.lag 0.75859 2.13527 0.26008 0.22146 3.43
log_dias_treat_imp_sin_na.lag 0.01313 1.01322 0.14279 0.12222 0.11
comp_bpsc_y3_severe.lag 0.10611 1.11194 0.17834 0.12291 0.86
policonsumo2.lag 0.02144 1.02167 0.22376 0.14550 0.15
edad_al_ing_1 -0.00175 0.99826 0.00887 0.00879 -0.20
ano_nac_corr -0.00275 0.99726 0.00859 0.00856 -0.32
susinidum_oh 0.05307 1.05451 0.15536 0.17220 0.31
susinidum_coc 0.15066 1.16261 0.18055 0.18971 0.79
susinidum_pbc 0.06123 1.06314 0.17256 0.19045 0.32
susinidum_mar 0.03864 1.03939 0.15764 0.17564 0.22
psycom_dum_with -0.02080 0.97941 0.04721 0.04682 -0.44
psycom_dum_study -0.06091 0.94091 0.06572 0.06445 -0.95
freq_cons_dum_5day -0.01579 0.98433 0.04386 0.04367 -0.36
cond_oc_dum_2inact 0.06525 1.06743 0.06486 0.06484 1.01
cond_oc_dum_3unemp 0.07212 1.07479 0.04734 0.04719 1.53
susprindum_oh 0.08099 1.08436 0.17644 0.18468 0.44
susprindum_coc 0.04766 1.04881 0.17743 0.18575 0.26
susprindum_pbc 0.07136 1.07397 0.17324 0.18145 0.39
susprindum_mar 0.13351 1.14283 0.19634 0.20052 0.67
Pr(>|z|)
tr_outcome.lag 0.15655
less_90d_tr1.lag 0.00061 ***
log_dias_treat_imp_sin_na.lag 0.91444
comp_bpsc_y3_severe.lag 0.38798
policonsumo2.lag 0.88285
edad_al_ing_1 0.84261
ano_nac_corr 0.74832
susinidum_oh 0.75792
susinidum_coc 0.42708
susinidum_pbc 0.74785
susinidum_mar 0.82589
psycom_dum_with 0.65677
psycom_dum_study 0.34465
freq_cons_dum_5day 0.71767
cond_oc_dum_2inact 0.31424
cond_oc_dum_3unemp 0.12643
susprindum_oh 0.66098
susprindum_coc 0.79751
susprindum_pbc 0.69410
susprindum_mar 0.50555
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.804 1.244 0.594 1.09
less_90d_tr1.lag 2.135 0.468 1.383 3.30
log_dias_treat_imp_sin_na.lag 1.013 0.987 0.797 1.29
comp_bpsc_y3_severe.lag 1.112 0.899 0.874 1.41
policonsumo2.lag 1.022 0.979 0.768 1.36
edad_al_ing_1 0.998 1.002 0.981 1.02
ano_nac_corr 0.997 1.003 0.981 1.01
susinidum_oh 1.055 0.948 0.752 1.48
susinidum_coc 1.163 0.860 0.802 1.69
susinidum_pbc 1.063 0.941 0.732 1.54
susinidum_mar 1.039 0.962 0.737 1.47
psycom_dum_with 0.979 1.021 0.894 1.07
psycom_dum_study 0.941 1.063 0.829 1.07
freq_cons_dum_5day 0.984 1.016 0.904 1.07
cond_oc_dum_2inact 1.067 0.937 0.940 1.21
cond_oc_dum_3unemp 1.075 0.930 0.980 1.18
susprindum_oh 1.084 0.922 0.755 1.56
susprindum_coc 1.049 0.953 0.729 1.51
susprindum_pbc 1.074 0.931 0.753 1.53
susprindum_mar 1.143 0.875 0.771 1.69
Concordance= 0.528 (se = 0.007 )
Likelihood ratio test= 67.3 on 20 df, p=5e-07
Wald test = 93.3 on 20 df, p=2e-11
Score (logrank) test = 57.5 on 20 df, p=2e-05, Robust = 47.3 p=5e-04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.365 0.433 0.450 0.471 0.466 1.154
[1] "Modality: basic ambulatory\nTime interval: (10,20]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1787, number of events= 1787
(1705 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.11289 0.89325 0.26499 0.19738 -0.57
less_90d_tr1.lag 0.30966 1.36296 0.42422 0.33394 0.93
log_dias_treat_imp_sin_na.lag -0.03569 0.96494 0.22487 0.17960 -0.20
comp_bpsc_y3_severe.lag -0.01927 0.98091 0.27409 0.21820 -0.09
policonsumo2.lag 0.30523 1.35693 0.23823 0.21449 1.42
edad_al_ing_1 -0.00444 0.99557 0.01000 0.00997 -0.45
ano_nac_corr -0.00873 0.99130 0.00971 0.00979 -0.89
susinidum_oh -0.10940 0.89637 0.17579 0.19112 -0.57
susinidum_coc -0.25157 0.77758 0.20960 0.21688 -1.16
susinidum_pbc -0.03937 0.96140 0.19588 0.21149 -0.19
susinidum_mar -0.07741 0.92551 0.17892 0.19344 -0.40
psycom_dum_with -0.05845 0.94322 0.05253 0.05173 -1.13
psycom_dum_study -0.14794 0.86249 0.07226 0.07361 -2.01
freq_cons_dum_5day 0.02945 1.02989 0.05516 0.05564 0.53
cond_oc_dum_2inact -0.02692 0.97344 0.07140 0.07258 -0.37
cond_oc_dum_3unemp -0.02013 0.98007 0.05935 0.05786 -0.35
susprindum_oh 0.04298 1.04392 0.20176 0.16993 0.25
susprindum_coc 0.05952 1.06133 0.20562 0.17324 0.34
susprindum_pbc -0.01571 0.98442 0.20156 0.17115 -0.09
susprindum_mar 0.22047 1.24666 0.21393 0.18470 1.19
Pr(>|z|)
tr_outcome.lag 0.567
less_90d_tr1.lag 0.354
log_dias_treat_imp_sin_na.lag 0.842
comp_bpsc_y3_severe.lag 0.930
policonsumo2.lag 0.155
edad_al_ing_1 0.656
ano_nac_corr 0.372
susinidum_oh 0.567
susinidum_coc 0.246
susinidum_pbc 0.852
susinidum_mar 0.689
psycom_dum_with 0.258
psycom_dum_study 0.044 *
freq_cons_dum_5day 0.597
cond_oc_dum_2inact 0.711
cond_oc_dum_3unemp 0.728
susprindum_oh 0.800
susprindum_coc 0.731
susprindum_pbc 0.927
susprindum_mar 0.233
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.893 1.120 0.607 1.315
less_90d_tr1.lag 1.363 0.734 0.708 2.623
log_dias_treat_imp_sin_na.lag 0.965 1.036 0.679 1.372
comp_bpsc_y3_severe.lag 0.981 1.019 0.640 1.504
policonsumo2.lag 1.357 0.737 0.891 2.066
edad_al_ing_1 0.996 1.004 0.976 1.015
ano_nac_corr 0.991 1.009 0.972 1.011
susinidum_oh 0.896 1.116 0.616 1.304
susinidum_coc 0.778 1.286 0.508 1.189
susinidum_pbc 0.961 1.040 0.635 1.455
susinidum_mar 0.926 1.080 0.633 1.352
psycom_dum_with 0.943 1.060 0.852 1.044
psycom_dum_study 0.862 1.159 0.747 0.996
freq_cons_dum_5day 1.030 0.971 0.923 1.149
cond_oc_dum_2inact 0.973 1.027 0.844 1.122
cond_oc_dum_3unemp 0.980 1.020 0.875 1.098
susprindum_oh 1.044 0.958 0.748 1.456
susprindum_coc 1.061 0.942 0.756 1.490
susprindum_pbc 0.984 1.016 0.704 1.377
susprindum_mar 1.247 0.802 0.868 1.790
Concordance= 0.535 (se = 0.008 )
Likelihood ratio test= 28.9 on 20 df, p=0.09
Wald test = 34.3 on 20 df, p=0.02
Score (logrank) test = 27.8 on 20 df, p=0.1, Robust = 29.2 p=0.08
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.473 0.647 0.688 0.701 0.737 1.481
[1] "Modality: GP residential\nTime interval: (10,20]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1104, number of events= 1104
(1012 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.20849 1.23182 0.22524 0.16977 1.23
less_90d_tr1.lag 0.18363 1.20157 0.28968 0.20840 0.88
log_dias_treat_imp_sin_na.lag -0.15939 0.85267 0.15317 0.09437 -1.69
comp_bpsc_y3_severe.lag 0.41048 1.50754 0.24543 0.19143 2.14
policonsumo2.lag 0.32097 1.37847 0.35877 0.20629 1.56
edad_al_ing_1 0.00535 1.00537 0.01386 0.01396 0.38
ano_nac_corr 0.00559 1.00561 0.01396 0.01365 0.41
susinidum_oh -0.15489 0.85651 0.21234 0.18000 -0.86
susinidum_coc -0.02269 0.97757 0.26191 0.22529 -0.10
susinidum_pbc -0.29053 0.74786 0.22752 0.19736 -1.47
susinidum_mar -0.14415 0.86575 0.21119 0.17822 -0.81
psycom_dum_with -0.07780 0.92515 0.07047 0.06905 -1.13
psycom_dum_study -0.18985 0.82708 0.09881 0.09247 -2.05
freq_cons_dum_5day -0.02728 0.97309 0.06870 0.06836 -0.40
cond_oc_dum_2inact -0.00107 0.99893 0.11269 0.11224 -0.01
cond_oc_dum_3unemp 0.11521 1.12211 0.09200 0.08840 1.30
susprindum_oh -0.04029 0.96051 0.31288 0.27695 -0.15
susprindum_coc -0.16429 0.84849 0.31546 0.28015 -0.59
susprindum_pbc -0.15457 0.85678 0.30654 0.27261 -0.57
susprindum_mar 0.41564 1.51533 0.35411 0.32712 1.27
Pr(>|z|)
tr_outcome.lag 0.219
less_90d_tr1.lag 0.378
log_dias_treat_imp_sin_na.lag 0.091 .
comp_bpsc_y3_severe.lag 0.032 *
policonsumo2.lag 0.120
edad_al_ing_1 0.701
ano_nac_corr 0.682
susinidum_oh 0.390
susinidum_coc 0.920
susinidum_pbc 0.141
susinidum_mar 0.419
psycom_dum_with 0.260
psycom_dum_study 0.040 *
freq_cons_dum_5day 0.690
cond_oc_dum_2inact 0.992
cond_oc_dum_3unemp 0.192
susprindum_oh 0.884
susprindum_coc 0.558
susprindum_pbc 0.571
susprindum_mar 0.204
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.232 0.812 0.883 1.718
less_90d_tr1.lag 1.202 0.832 0.799 1.808
log_dias_treat_imp_sin_na.lag 0.853 1.173 0.709 1.026
comp_bpsc_y3_severe.lag 1.508 0.663 1.036 2.194
policonsumo2.lag 1.378 0.725 0.920 2.065
edad_al_ing_1 1.005 0.995 0.978 1.033
ano_nac_corr 1.006 0.994 0.979 1.033
susinidum_oh 0.857 1.168 0.602 1.219
susinidum_coc 0.978 1.023 0.629 1.520
susinidum_pbc 0.748 1.337 0.508 1.101
susinidum_mar 0.866 1.155 0.611 1.228
psycom_dum_with 0.925 1.081 0.808 1.059
psycom_dum_study 0.827 1.209 0.690 0.991
freq_cons_dum_5day 0.973 1.028 0.851 1.113
cond_oc_dum_2inact 0.999 1.001 0.802 1.245
cond_oc_dum_3unemp 1.122 0.891 0.944 1.334
susprindum_oh 0.961 1.041 0.558 1.653
susprindum_coc 0.848 1.179 0.490 1.469
susprindum_pbc 0.857 1.167 0.502 1.462
susprindum_mar 1.515 0.660 0.798 2.877
Concordance= 0.549 (se = 0.01 )
Likelihood ratio test= 50.8 on 20 df, p=2e-04
Wald test = 69.5 on 20 df, p=2e-07
Score (logrank) test = 47.6 on 20 df, p=5e-04, Robust = 46.6 p=7e-04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.194 0.403 0.432 0.459 0.478 1.500
[1] "Modality: WO intensive ambulatory\nTime interval: (10,20]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 383, number of events= 383
(363 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.0634 1.0654 0.5857 0.2750 0.23
less_90d_tr1.lag 0.1451 1.1562 0.6219 0.2496 0.58
log_dias_treat_imp_sin_na.lag -0.0925 0.9116 0.4925 0.2183 -0.42
comp_bpsc_y3_severe.lag 0.2509 1.2852 0.4928 0.2624 0.96
policonsumo2.lag 0.1765 1.1931 0.5414 0.2508 0.70
edad_al_ing_1 0.0183 1.0185 0.0229 0.0233 0.79
ano_nac_corr 0.0113 1.0113 0.0226 0.0217 0.52
susinidum_oh -0.2012 0.8178 0.3296 0.2297 -0.88
susinidum_coc -0.2162 0.8055 0.3882 0.2877 -0.75
susinidum_pbc -0.3864 0.6795 0.3519 0.2679 -1.44
susinidum_mar -0.2634 0.7684 0.3302 0.2266 -1.16
psycom_dum_with -0.0683 0.9340 0.1235 0.1174 -0.58
psycom_dum_study -0.0856 0.9179 0.1662 0.1586 -0.54
freq_cons_dum_5day 0.0226 1.0229 0.1079 0.1064 0.21
cond_oc_dum_2inact -0.2520 0.7773 0.1403 0.1450 -1.74
cond_oc_dum_3unemp -0.1483 0.8621 0.1438 0.1474 -1.01
susprindum_oh -0.1126 0.8935 0.4137 0.3334 -0.34
susprindum_coc 0.1187 1.1260 0.4111 0.3188 0.37
susprindum_pbc 0.0363 1.0370 0.4006 0.3042 0.12
susprindum_mar 0.3433 1.4095 0.4282 0.3499 0.98
Pr(>|z|)
tr_outcome.lag 0.818
less_90d_tr1.lag 0.561
log_dias_treat_imp_sin_na.lag 0.672
comp_bpsc_y3_severe.lag 0.339
policonsumo2.lag 0.482
edad_al_ing_1 0.431
ano_nac_corr 0.604
susinidum_oh 0.381
susinidum_coc 0.452
susinidum_pbc 0.149
susinidum_mar 0.245
psycom_dum_with 0.561
psycom_dum_study 0.589
freq_cons_dum_5day 0.832
cond_oc_dum_2inact 0.082 .
cond_oc_dum_3unemp 0.314
susprindum_oh 0.736
susprindum_coc 0.710
susprindum_pbc 0.905
susprindum_mar 0.327
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.065 0.939 0.622 1.83
less_90d_tr1.lag 1.156 0.865 0.709 1.89
log_dias_treat_imp_sin_na.lag 0.912 1.097 0.594 1.40
comp_bpsc_y3_severe.lag 1.285 0.778 0.768 2.15
policonsumo2.lag 1.193 0.838 0.730 1.95
edad_al_ing_1 1.018 0.982 0.973 1.07
ano_nac_corr 1.011 0.989 0.969 1.06
susinidum_oh 0.818 1.223 0.521 1.28
susinidum_coc 0.806 1.241 0.458 1.42
susinidum_pbc 0.680 1.472 0.402 1.15
susinidum_mar 0.768 1.301 0.493 1.20
psycom_dum_with 0.934 1.071 0.742 1.18
psycom_dum_study 0.918 1.089 0.673 1.25
freq_cons_dum_5day 1.023 0.978 0.830 1.26
cond_oc_dum_2inact 0.777 1.287 0.585 1.03
cond_oc_dum_3unemp 0.862 1.160 0.646 1.15
susprindum_oh 0.894 1.119 0.465 1.72
susprindum_coc 1.126 0.888 0.603 2.10
susprindum_pbc 1.037 0.964 0.571 1.88
susprindum_mar 1.410 0.709 0.710 2.80
Concordance= 0.558 (se = 0.016 )
Likelihood ratio test= 14.6 on 20 df, p=0.8
Wald test = 22.8 on 20 df, p=0.3
Score (logrank) test = 14.9 on 20 df, p=0.8, Robust = 21.6 p=0.4
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.401 0.701 0.802 0.808 0.907 1.542
[1] "Modality: WO residential\nTime interval: (20,30]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 337, number of events= 337
(312 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.54805 0.57808 0.63302 0.56170 -0.98
less_90d_tr1.lag 1.55245 4.72302 0.44287 0.48428 3.21
log_dias_treat_imp_sin_na.lag 0.44260 1.55674 0.23419 0.18393 2.41
comp_bpsc_y3_severe.lag -0.33788 0.71328 0.54061 0.46918 -0.72
policonsumo2.lag -1.38744 0.24971 0.65732 0.48749 -2.85
edad_al_ing_1 0.00518 1.00520 0.02817 0.02923 0.18
ano_nac_corr 0.01173 1.01180 0.02745 0.02832 0.41
susinidum_oh -0.89518 0.40854 0.36761 0.34203 -2.62
susinidum_coc -0.42191 0.65579 0.44454 0.46681 -0.90
susinidum_pbc -0.80592 0.44668 0.38239 0.35110 -2.30
susinidum_mar -0.67735 0.50796 0.37139 0.34057 -1.99
psycom_dum_with 0.11314 1.11979 0.16892 0.16214 0.70
psycom_dum_study 0.17197 1.18764 0.19648 0.18875 0.91
freq_cons_dum_5day -0.12210 0.88506 0.13524 0.12778 -0.96
cond_oc_dum_2inact 0.01466 1.01477 0.21303 0.20392 0.07
cond_oc_dum_3unemp 0.11172 1.11820 0.21404 0.20170 0.55
susprindum_oh 0.60878 1.83818 0.36597 0.33412 1.82
susprindum_coc 0.20307 1.22516 0.36591 0.34566 0.59
susprindum_pbc 0.55324 1.73888 0.34393 0.32150 1.72
susprindum_mar -0.06475 0.93731 0.49574 0.46418 -0.14
Pr(>|z|)
tr_outcome.lag 0.3292
less_90d_tr1.lag 0.0013 **
log_dias_treat_imp_sin_na.lag 0.0161 *
comp_bpsc_y3_severe.lag 0.4714
policonsumo2.lag 0.0044 **
edad_al_ing_1 0.8593
ano_nac_corr 0.6788
susinidum_oh 0.0089 **
susinidum_coc 0.3661
susinidum_pbc 0.0217 *
susinidum_mar 0.0467 *
psycom_dum_with 0.4853
psycom_dum_study 0.3622
freq_cons_dum_5day 0.3393
cond_oc_dum_2inact 0.9427
cond_oc_dum_3unemp 0.5796
susprindum_oh 0.0685 .
susprindum_coc 0.5569
susprindum_pbc 0.0853 .
susprindum_mar 0.8891
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.578 1.730 0.192 1.738
less_90d_tr1.lag 4.723 0.212 1.828 12.202
log_dias_treat_imp_sin_na.lag 1.557 0.642 1.086 2.232
comp_bpsc_y3_severe.lag 0.713 1.402 0.284 1.789
policonsumo2.lag 0.250 4.005 0.096 0.649
edad_al_ing_1 1.005 0.995 0.949 1.064
ano_nac_corr 1.012 0.988 0.957 1.070
susinidum_oh 0.409 2.448 0.209 0.799
susinidum_coc 0.656 1.525 0.263 1.637
susinidum_pbc 0.447 2.239 0.224 0.889
susinidum_mar 0.508 1.969 0.261 0.990
psycom_dum_with 1.120 0.893 0.815 1.539
psycom_dum_study 1.188 0.842 0.820 1.719
freq_cons_dum_5day 0.885 1.130 0.689 1.137
cond_oc_dum_2inact 1.015 0.985 0.680 1.513
cond_oc_dum_3unemp 1.118 0.894 0.753 1.660
susprindum_oh 1.838 0.544 0.955 3.538
susprindum_coc 1.225 0.816 0.622 2.412
susprindum_pbc 1.739 0.575 0.926 3.265
susprindum_mar 0.937 1.067 0.377 2.328
Concordance= 0.58 (se = 0.019 )
Likelihood ratio test= 33.7 on 20 df, p=0.03
Wald test = 30.1 on 20 df, p=0.07
Score (logrank) test = 32.4 on 20 df, p=0.04, Robust = 29.8 p=0.07
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.94 2.20 2.54 2.79 2.96 11.34
[1] "Modality: GP intensive ambulatory\nTime interval: (20,30]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1468, number of events= 1468
(1416 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.88258 0.41371 0.34605 0.23629 -3.74
less_90d_tr1.lag 0.69646 2.00664 0.30727 0.18314 3.80
log_dias_treat_imp_sin_na.lag -0.11098 0.89496 0.17582 0.11884 -0.93
comp_bpsc_y3_severe.lag 0.09737 1.10227 0.26731 0.19971 0.49
policonsumo2.lag -0.16274 0.84981 0.37994 0.19389 -0.84
edad_al_ing_1 0.00906 1.00911 0.01199 0.01156 0.78
ano_nac_corr 0.01241 1.01249 0.01185 0.01152 1.08
susinidum_oh -0.17156 0.84235 0.17261 0.14429 -1.19
susinidum_coc -0.12936 0.87865 0.20917 0.18547 -0.70
susinidum_pbc -0.24760 0.78068 0.19361 0.16656 -1.49
susinidum_mar -0.14856 0.86195 0.17518 0.14781 -1.01
psycom_dum_with 0.07155 1.07417 0.05849 0.05756 1.24
psycom_dum_study 0.13765 1.14757 0.07960 0.08249 1.67
freq_cons_dum_5day -0.02325 0.97702 0.05463 0.05461 -0.43
cond_oc_dum_2inact -0.16820 0.84519 0.08031 0.07533 -2.23
cond_oc_dum_3unemp -0.07700 0.92589 0.05856 0.05873 -1.31
susprindum_oh 0.34467 1.41152 0.22480 0.22376 1.54
susprindum_coc 0.37247 1.45131 0.22768 0.22788 1.63
susprindum_pbc 0.36381 1.43880 0.21944 0.21983 1.65
susprindum_mar 0.54116 1.71799 0.24309 0.24170 2.24
Pr(>|z|)
tr_outcome.lag 0.00019 ***
less_90d_tr1.lag 0.00014 ***
log_dias_treat_imp_sin_na.lag 0.35037
comp_bpsc_y3_severe.lag 0.62586
policonsumo2.lag 0.40130
edad_al_ing_1 0.43300
ano_nac_corr 0.28143
susinidum_oh 0.23445
susinidum_coc 0.48548
susinidum_pbc 0.13714
susinidum_mar 0.31485
psycom_dum_with 0.21384
psycom_dum_study 0.09516 .
freq_cons_dum_5day 0.67023
cond_oc_dum_2inact 0.02556 *
cond_oc_dum_3unemp 0.18983
susprindum_oh 0.12347
susprindum_coc 0.10216
susprindum_pbc 0.09793 .
susprindum_mar 0.02516 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.414 2.417 0.260 0.657
less_90d_tr1.lag 2.007 0.498 1.401 2.873
log_dias_treat_imp_sin_na.lag 0.895 1.117 0.709 1.130
comp_bpsc_y3_severe.lag 1.102 0.907 0.745 1.630
policonsumo2.lag 0.850 1.177 0.581 1.243
edad_al_ing_1 1.009 0.991 0.986 1.032
ano_nac_corr 1.012 0.988 0.990 1.036
susinidum_oh 0.842 1.187 0.635 1.118
susinidum_coc 0.879 1.138 0.611 1.264
susinidum_pbc 0.781 1.281 0.563 1.082
susinidum_mar 0.862 1.160 0.645 1.152
psycom_dum_with 1.074 0.931 0.960 1.202
psycom_dum_study 1.148 0.871 0.976 1.349
freq_cons_dum_5day 0.977 1.024 0.878 1.087
cond_oc_dum_2inact 0.845 1.183 0.729 0.980
cond_oc_dum_3unemp 0.926 1.080 0.825 1.039
susprindum_oh 1.412 0.708 0.910 2.189
susprindum_coc 1.451 0.689 0.929 2.268
susprindum_pbc 1.439 0.695 0.935 2.214
susprindum_mar 1.718 0.582 1.070 2.759
Concordance= 0.536 (se = 0.009 )
Likelihood ratio test= 46.4 on 20 df, p=7e-04
Wald test = 54 on 20 df, p=6e-05
Score (logrank) test = 41.5 on 20 df, p=0.003, Robust = 35.7 p=0.02
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.662 0.987 1.058 1.093 1.145 3.159
[1] "Modality: basic ambulatory\nTime interval: (20,30]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1208, number of events= 1208
(1172 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.2222 1.2488 0.4500 0.3032 0.73
less_90d_tr1.lag 0.8375 2.3106 0.5200 0.4845 1.73
log_dias_treat_imp_sin_na.lag 0.4142 1.5132 0.3048 0.2076 1.99
comp_bpsc_y3_severe.lag 0.0768 1.0799 0.3167 0.3050 0.25
policonsumo2.lag 0.3782 1.4596 0.4335 0.3065 1.23
edad_al_ing_1 0.0414 1.0422 0.0138 0.0132 3.14
ano_nac_corr 0.0402 1.0410 0.0136 0.0129 3.12
susinidum_oh 0.0310 1.0315 0.2196 0.2160 0.14
susinidum_coc 0.0954 1.1001 0.2557 0.2466 0.39
susinidum_pbc 0.2317 1.2608 0.2497 0.2416 0.96
susinidum_mar 0.0550 1.0566 0.2215 0.2192 0.25
psycom_dum_with -0.0615 0.9403 0.0666 0.0678 -0.91
psycom_dum_study -0.1058 0.8996 0.0799 0.0743 -1.42
freq_cons_dum_5day 0.0800 1.0832 0.0674 0.0653 1.22
cond_oc_dum_2inact -0.0123 0.9878 0.0959 0.0935 -0.13
cond_oc_dum_3unemp -0.0226 0.9777 0.0695 0.0705 -0.32
susprindum_oh -0.4384 0.6451 0.2856 0.2298 -1.91
susprindum_coc -0.4450 0.6408 0.2898 0.2350 -1.89
susprindum_pbc -0.4805 0.6185 0.2855 0.2282 -2.11
susprindum_mar -0.4434 0.6418 0.2988 0.2540 -1.75
Pr(>|z|)
tr_outcome.lag 0.4637
less_90d_tr1.lag 0.0839 .
log_dias_treat_imp_sin_na.lag 0.0460 *
comp_bpsc_y3_severe.lag 0.8011
policonsumo2.lag 0.2172
edad_al_ing_1 0.0017 **
ano_nac_corr 0.0018 **
susinidum_oh 0.8859
susinidum_coc 0.6989
susinidum_pbc 0.3374
susinidum_mar 0.8018
psycom_dum_with 0.3641
psycom_dum_study 0.1543
freq_cons_dum_5day 0.2208
cond_oc_dum_2inact 0.8957
cond_oc_dum_3unemp 0.7489
susprindum_oh 0.0564 .
susprindum_coc 0.0582 .
susprindum_pbc 0.0352 *
susprindum_mar 0.0809 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.249 0.801 0.689 2.263
less_90d_tr1.lag 2.311 0.433 0.894 5.972
log_dias_treat_imp_sin_na.lag 1.513 0.661 1.007 2.273
comp_bpsc_y3_severe.lag 1.080 0.926 0.594 1.963
policonsumo2.lag 1.460 0.685 0.800 2.661
edad_al_ing_1 1.042 0.959 1.016 1.069
ano_nac_corr 1.041 0.961 1.015 1.068
susinidum_oh 1.031 0.969 0.675 1.575
susinidum_coc 1.100 0.909 0.678 1.784
susinidum_pbc 1.261 0.793 0.785 2.024
susinidum_mar 1.057 0.946 0.688 1.623
psycom_dum_with 0.940 1.063 0.823 1.074
psycom_dum_study 0.900 1.112 0.778 1.041
freq_cons_dum_5day 1.083 0.923 0.953 1.231
cond_oc_dum_2inact 0.988 1.012 0.822 1.187
cond_oc_dum_3unemp 0.978 1.023 0.851 1.123
susprindum_oh 0.645 1.550 0.411 1.012
susprindum_coc 0.641 1.561 0.404 1.016
susprindum_pbc 0.618 1.617 0.395 0.967
susprindum_mar 0.642 1.558 0.390 1.056
Concordance= 0.552 (se = 0.009 )
Likelihood ratio test= 29 on 20 df, p=0.09
Wald test = 43.1 on 20 df, p=0.002
Score (logrank) test = 27.6 on 20 df, p=0.1, Robust = 32 p=0.04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.201 0.314 0.343 0.349 0.374 1.062
[1] "Modality: GP residential\nTime interval: (20,30]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 612, number of events= 612
(578 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.0126 1.0127 0.5116 0.2984 0.04
less_90d_tr1.lag 0.3437 1.4102 0.4858 0.3058 1.12
log_dias_treat_imp_sin_na.lag -0.1050 0.9003 0.2077 0.1210 -0.87
comp_bpsc_y3_severe.lag 0.4596 1.5835 0.4627 0.2917 1.58
policonsumo2.lag 0.5207 1.6832 0.5941 0.3723 1.40
edad_al_ing_1 0.0165 1.0167 0.0203 0.0182 0.91
ano_nac_corr 0.0213 1.0215 0.0200 0.0183 1.16
susinidum_oh -0.4668 0.6270 0.3163 0.2554 -1.83
susinidum_coc -0.0377 0.9630 0.3617 0.2968 -0.13
susinidum_pbc -0.2989 0.7416 0.3315 0.2661 -1.12
susinidum_mar -0.4887 0.6134 0.3163 0.2526 -1.94
psycom_dum_with -0.1009 0.9040 0.0943 0.0976 -1.03
psycom_dum_study -0.0178 0.9824 0.1338 0.1232 -0.14
freq_cons_dum_5day 0.1393 1.1494 0.0949 0.0940 1.48
cond_oc_dum_2inact -0.0755 0.9272 0.1501 0.1472 -0.51
cond_oc_dum_3unemp -0.0611 0.9407 0.1294 0.1260 -0.49
susprindum_oh 0.8357 2.3065 0.4654 0.4046 2.07
susprindum_coc 0.8076 2.2426 0.4702 0.4086 1.98
susprindum_pbc 0.8110 2.2503 0.4570 0.3925 2.07
susprindum_mar 0.5686 1.7657 0.5253 0.4503 1.26
Pr(>|z|)
tr_outcome.lag 0.966
less_90d_tr1.lag 0.261
log_dias_treat_imp_sin_na.lag 0.386
comp_bpsc_y3_severe.lag 0.115
policonsumo2.lag 0.162
edad_al_ing_1 0.365
ano_nac_corr 0.246
susinidum_oh 0.068 .
susinidum_coc 0.899
susinidum_pbc 0.261
susinidum_mar 0.053 .
psycom_dum_with 0.301
psycom_dum_study 0.885
freq_cons_dum_5day 0.138
cond_oc_dum_2inact 0.608
cond_oc_dum_3unemp 0.628
susprindum_oh 0.039 *
susprindum_coc 0.048 *
susprindum_pbc 0.039 *
susprindum_mar 0.207
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.013 0.987 0.564 1.82
less_90d_tr1.lag 1.410 0.709 0.774 2.57
log_dias_treat_imp_sin_na.lag 0.900 1.111 0.710 1.14
comp_bpsc_y3_severe.lag 1.583 0.632 0.894 2.81
policonsumo2.lag 1.683 0.594 0.811 3.49
edad_al_ing_1 1.017 0.984 0.981 1.05
ano_nac_corr 1.022 0.979 0.985 1.06
susinidum_oh 0.627 1.595 0.380 1.03
susinidum_coc 0.963 1.038 0.538 1.72
susinidum_pbc 0.742 1.348 0.440 1.25
susinidum_mar 0.613 1.630 0.374 1.01
psycom_dum_with 0.904 1.106 0.747 1.09
psycom_dum_study 0.982 1.018 0.772 1.25
freq_cons_dum_5day 1.149 0.870 0.956 1.38
cond_oc_dum_2inact 0.927 1.078 0.695 1.24
cond_oc_dum_3unemp 0.941 1.063 0.735 1.20
susprindum_oh 2.307 0.434 1.044 5.10
susprindum_coc 2.243 0.446 1.007 5.00
susprindum_pbc 2.250 0.444 1.043 4.86
susprindum_mar 1.766 0.566 0.730 4.27
Concordance= 0.549 (se = 0.013 )
Likelihood ratio test= 30.4 on 20 df, p=0.06
Wald test = 47.8 on 20 df, p=5e-04
Score (logrank) test = 28.6 on 20 df, p=0.1, Robust = 28.3 p=0.1
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.089 0.169 0.184 0.191 0.201 0.559
[1] "Modality: WO intensive ambulatory\nTime interval: (20,30]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 247, number of events= 247
(235 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.40479 1.49898 1.04300 0.48419 0.84
less_90d_tr1.lag 0.08277 1.08629 0.91450 0.49716 0.17
log_dias_treat_imp_sin_na.lag 0.19472 1.21497 0.76258 0.39201 0.50
comp_bpsc_y3_severe.lag 0.43809 1.54974 0.87445 0.57240 0.77
policonsumo2.lag 0.97602 2.65387 0.77855 0.63265 1.54
edad_al_ing_1 -0.01578 0.98434 0.03117 0.02972 -0.53
ano_nac_corr -0.01833 0.98184 0.02929 0.02678 -0.68
susinidum_oh -0.54823 0.57797 0.55667 0.47227 -1.16
susinidum_coc -0.41163 0.66257 0.63174 0.54671 -0.75
susinidum_pbc -0.72558 0.48404 0.59230 0.50021 -1.45
susinidum_mar -0.44927 0.63809 0.56526 0.47497 -0.95
psycom_dum_with 0.02344 1.02371 0.14638 0.14068 0.17
psycom_dum_study 0.30425 1.35561 0.23596 0.25941 1.17
freq_cons_dum_5day -0.25259 0.77679 0.14603 0.13977 -1.81
cond_oc_dum_2inact -0.05708 0.94452 0.17429 0.15677 -0.36
cond_oc_dum_3unemp -0.00650 0.99352 0.18484 0.18418 -0.04
susprindum_oh -0.43071 0.65005 0.45904 0.35991 -1.20
susprindum_coc 0.12234 1.13014 0.47224 0.39638 0.31
susprindum_pbc -0.00815 0.99189 0.46215 0.37515 -0.02
susprindum_mar -0.01805 0.98211 0.51912 0.40588 -0.04
Pr(>|z|)
tr_outcome.lag 0.403
less_90d_tr1.lag 0.868
log_dias_treat_imp_sin_na.lag 0.619
comp_bpsc_y3_severe.lag 0.444
policonsumo2.lag 0.123
edad_al_ing_1 0.595
ano_nac_corr 0.494
susinidum_oh 0.246
susinidum_coc 0.452
susinidum_pbc 0.147
susinidum_mar 0.344
psycom_dum_with 0.868
psycom_dum_study 0.241
freq_cons_dum_5day 0.071 .
cond_oc_dum_2inact 0.716
cond_oc_dum_3unemp 0.972
susprindum_oh 0.231
susprindum_coc 0.758
susprindum_pbc 0.983
susprindum_mar 0.965
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.499 0.667 0.580 3.87
less_90d_tr1.lag 1.086 0.921 0.410 2.88
log_dias_treat_imp_sin_na.lag 1.215 0.823 0.563 2.62
comp_bpsc_y3_severe.lag 1.550 0.645 0.505 4.76
policonsumo2.lag 2.654 0.377 0.768 9.17
edad_al_ing_1 0.984 1.016 0.929 1.04
ano_nac_corr 0.982 1.018 0.932 1.03
susinidum_oh 0.578 1.730 0.229 1.46
susinidum_coc 0.663 1.509 0.227 1.93
susinidum_pbc 0.484 2.066 0.182 1.29
susinidum_mar 0.638 1.567 0.252 1.62
psycom_dum_with 1.024 0.977 0.777 1.35
psycom_dum_study 1.356 0.738 0.815 2.25
freq_cons_dum_5day 0.777 1.287 0.591 1.02
cond_oc_dum_2inact 0.945 1.059 0.695 1.28
cond_oc_dum_3unemp 0.994 1.007 0.692 1.43
susprindum_oh 0.650 1.538 0.321 1.32
susprindum_coc 1.130 0.885 0.520 2.46
susprindum_pbc 0.992 1.008 0.475 2.07
susprindum_mar 0.982 1.018 0.443 2.18
Concordance= 0.589 (se = 0.021 )
Likelihood ratio test= 17.7 on 20 df, p=0.6
Wald test = 28.8 on 20 df, p=0.09
Score (logrank) test = 17.2 on 20 df, p=0.6, Robust = 21 p=0.4
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.143 0.236 0.286 0.305 0.360 1.005
[1] "Modality: WO residential\nTime interval: (30,50]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 374, number of events= 374
(332 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.1910 0.8261 0.3716 0.1834 -1.04
less_90d_tr1.lag 0.3183 1.3749 0.2421 0.1323 2.41
log_dias_treat_imp_sin_na.lag -0.0582 0.9434 0.0852 0.0303 -1.92
comp_bpsc_y3_severe.lag 0.2192 1.2451 0.4037 0.1521 1.44
policonsumo2.lag -0.1290 0.8789 0.5011 0.1685 -0.77
edad_al_ing_1 0.0669 1.0692 0.0311 0.0320 2.09
ano_nac_corr 0.0667 1.0689 0.0305 0.0315 2.12
susinidum_oh 1.1813 3.2587 0.5603 0.3102 3.81
susinidum_coc 1.4419 4.2287 0.6052 0.3416 4.22
susinidum_pbc 1.1537 3.1698 0.5694 0.3223 3.58
susinidum_mar 1.1777 3.2468 0.5616 0.3072 3.83
psycom_dum_with -0.0454 0.9557 0.1506 0.1451 -0.31
psycom_dum_study 0.0300 1.0305 0.1714 0.1609 0.19
freq_cons_dum_5day -0.0269 0.9734 0.1255 0.1198 -0.22
cond_oc_dum_2inact -0.1698 0.8439 0.2263 0.1750 -0.97
cond_oc_dum_3unemp -0.2563 0.7739 0.2304 0.1896 -1.35
susprindum_oh -0.8702 0.4189 0.3963 0.2483 -3.50
susprindum_coc -0.9727 0.3781 0.4019 0.2531 -3.84
susprindum_pbc -0.8575 0.4242 0.3773 0.2169 -3.95
susprindum_mar -0.4442 0.6413 0.5670 0.3156 -1.41
Pr(>|z|)
tr_outcome.lag 0.29773
less_90d_tr1.lag 0.01615 *
log_dias_treat_imp_sin_na.lag 0.05447 .
comp_bpsc_y3_severe.lag 0.14948
policonsumo2.lag 0.44375
edad_al_ing_1 0.03654 *
ano_nac_corr 0.03411 *
susinidum_oh 0.00014 ***
susinidum_coc 2.4e-05 ***
susinidum_pbc 0.00034 ***
susinidum_mar 0.00013 ***
psycom_dum_with 0.75448
psycom_dum_study 0.85194
freq_cons_dum_5day 0.82209
cond_oc_dum_2inact 0.33203
cond_oc_dum_3unemp 0.17642
susprindum_oh 0.00046 ***
susprindum_coc 0.00012 ***
susprindum_pbc 7.7e-05 ***
susprindum_mar 0.15934
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.826 1.210 0.577 1.184
less_90d_tr1.lag 1.375 0.727 1.061 1.782
log_dias_treat_imp_sin_na.lag 0.943 1.060 0.889 1.001
comp_bpsc_y3_severe.lag 1.245 0.803 0.924 1.677
policonsumo2.lag 0.879 1.138 0.632 1.223
edad_al_ing_1 1.069 0.935 1.004 1.138
ano_nac_corr 1.069 0.935 1.005 1.137
susinidum_oh 3.259 0.307 1.774 5.985
susinidum_coc 4.229 0.236 2.165 8.261
susinidum_pbc 3.170 0.315 1.685 5.962
susinidum_mar 3.247 0.308 1.778 5.928
psycom_dum_with 0.956 1.046 0.719 1.270
psycom_dum_study 1.030 0.970 0.752 1.413
freq_cons_dum_5day 0.973 1.027 0.770 1.231
cond_oc_dum_2inact 0.844 1.185 0.599 1.189
cond_oc_dum_3unemp 0.774 1.292 0.534 1.122
susprindum_oh 0.419 2.387 0.257 0.681
susprindum_coc 0.378 2.645 0.230 0.621
susprindum_pbc 0.424 2.357 0.277 0.649
susprindum_mar 0.641 1.559 0.345 1.191
Concordance= 0.558 (se = 0.016 )
Likelihood ratio test= 21.5 on 20 df, p=0.4
Wald test = 72.6 on 20 df, p=7e-08
Score (logrank) test = 20.9 on 20 df, p=0.4, Robust = 31.6 p=0.05
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.245 0.662 0.747 0.776 0.873 2.800
[1] "Modality: GP intensive ambulatory\nTime interval: (30,50]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1813, number of events= 1813
(1667 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.14332 1.15409 0.22733 0.15475 0.93
less_90d_tr1.lag 0.07829 1.08144 0.15134 0.13201 0.59
log_dias_treat_imp_sin_na.lag -0.07846 0.92454 0.06789 0.02680 -2.93
comp_bpsc_y3_severe.lag 0.22928 1.25770 0.14756 0.12873 1.78
policonsumo2.lag -0.04683 0.95425 0.22188 0.15928 -0.29
edad_al_ing_1 0.04965 1.05091 0.01264 0.01231 4.03
ano_nac_corr 0.05133 1.05267 0.01264 0.01252 4.10
susinidum_oh -0.22128 0.80149 0.18650 0.15174 -1.46
susinidum_coc -0.10405 0.90118 0.21627 0.18393 -0.57
susinidum_pbc -0.29351 0.74564 0.20573 0.17919 -1.64
susinidum_mar -0.26823 0.76473 0.18907 0.15569 -1.72
psycom_dum_with 0.03134 1.03183 0.05337 0.05347 0.59
psycom_dum_study -0.02588 0.97445 0.07087 0.07111 -0.36
freq_cons_dum_5day 0.02118 1.02141 0.04865 0.04795 0.44
cond_oc_dum_2inact -0.00426 0.99575 0.07192 0.07471 -0.06
cond_oc_dum_3unemp 0.09166 1.09600 0.05231 0.05118 1.79
susprindum_oh 0.02147 1.02171 0.21852 0.17270 0.12
susprindum_coc 0.07592 1.07887 0.22263 0.17813 0.43
susprindum_pbc 0.05355 1.05501 0.21734 0.17184 0.31
susprindum_mar 0.05373 1.05520 0.24219 0.19594 0.27
Pr(>|z|)
tr_outcome.lag 0.3544
less_90d_tr1.lag 0.5531
log_dias_treat_imp_sin_na.lag 0.0034 **
comp_bpsc_y3_severe.lag 0.0749 .
policonsumo2.lag 0.7687
edad_al_ing_1 5.5e-05 ***
ano_nac_corr 4.1e-05 ***
susinidum_oh 0.1448
susinidum_coc 0.5716
susinidum_pbc 0.1014
susinidum_mar 0.0849 .
psycom_dum_with 0.5579
psycom_dum_study 0.7159
freq_cons_dum_5day 0.6586
cond_oc_dum_2inact 0.9545
cond_oc_dum_3unemp 0.0733 .
susprindum_oh 0.9011
susprindum_coc 0.6700
susprindum_pbc 0.7553
susprindum_mar 0.7839
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.154 0.866 0.852 1.563
less_90d_tr1.lag 1.081 0.925 0.835 1.401
log_dias_treat_imp_sin_na.lag 0.925 1.082 0.877 0.974
comp_bpsc_y3_severe.lag 1.258 0.795 0.977 1.619
policonsumo2.lag 0.954 1.048 0.698 1.304
edad_al_ing_1 1.051 0.952 1.026 1.077
ano_nac_corr 1.053 0.950 1.027 1.079
susinidum_oh 0.801 1.248 0.595 1.079
susinidum_coc 0.901 1.110 0.628 1.292
susinidum_pbc 0.746 1.341 0.525 1.059
susinidum_mar 0.765 1.308 0.564 1.038
psycom_dum_with 1.032 0.969 0.929 1.146
psycom_dum_study 0.974 1.026 0.848 1.120
freq_cons_dum_5day 1.021 0.979 0.930 1.122
cond_oc_dum_2inact 0.996 1.004 0.860 1.153
cond_oc_dum_3unemp 1.096 0.912 0.991 1.212
susprindum_oh 1.022 0.979 0.728 1.433
susprindum_coc 1.079 0.927 0.761 1.530
susprindum_pbc 1.055 0.948 0.753 1.478
susprindum_mar 1.055 0.948 0.719 1.549
Concordance= 0.541 (se = 0.008 )
Likelihood ratio test= 38.2 on 20 df, p=0.008
Wald test = 52.9 on 20 df, p=8e-05
Score (logrank) test = 37.1 on 20 df, p=0.01, Robust = 47.9 p=4e-04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.467 0.699 0.763 0.780 0.840 1.641
[1] "Modality: basic ambulatory\nTime interval: (30,50]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 1568, number of events= 1568
(1465 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.2396 0.7869 0.2605 0.1952 -1.23
less_90d_tr1.lag 0.1666 1.1813 0.2345 0.2055 0.81
log_dias_treat_imp_sin_na.lag 0.0305 1.0310 0.1362 0.1365 0.22
comp_bpsc_y3_severe.lag 0.1990 1.2201 0.2024 0.1578 1.26
policonsumo2.lag 0.2099 1.2336 0.2926 0.2411 0.87
edad_al_ing_1 0.0621 1.0640 0.0137 0.0134 4.64
ano_nac_corr 0.0617 1.0636 0.0134 0.0134 4.60
susinidum_oh 0.2714 1.3117 0.1894 0.1984 1.37
susinidum_coc 0.1121 1.1186 0.2247 0.2401 0.47
susinidum_pbc 0.2334 1.2629 0.2212 0.2241 1.04
susinidum_mar 0.3088 1.3618 0.1922 0.2027 1.52
psycom_dum_with 0.0789 1.0821 0.0567 0.0574 1.38
psycom_dum_study 0.0741 1.0769 0.0745 0.0720 1.03
freq_cons_dum_5day -0.0672 0.9350 0.0585 0.0580 -1.16
cond_oc_dum_2inact -0.0729 0.9297 0.0799 0.0772 -0.94
cond_oc_dum_3unemp 0.0917 1.0960 0.0615 0.0612 1.50
susprindum_oh -0.1605 0.8517 0.2267 0.1864 -0.86
susprindum_coc -0.1131 0.8930 0.2313 0.1932 -0.59
susprindum_pbc -0.1756 0.8390 0.2260 0.1865 -0.94
susprindum_mar -0.2675 0.7653 0.2436 0.2004 -1.33
Pr(>|z|)
tr_outcome.lag 0.22
less_90d_tr1.lag 0.42
log_dias_treat_imp_sin_na.lag 0.82
comp_bpsc_y3_severe.lag 0.21
policonsumo2.lag 0.38
edad_al_ing_1 3.5e-06 ***
ano_nac_corr 4.1e-06 ***
susinidum_oh 0.17
susinidum_coc 0.64
susinidum_pbc 0.30
susinidum_mar 0.13
psycom_dum_with 0.17
psycom_dum_study 0.30
freq_cons_dum_5day 0.25
cond_oc_dum_2inact 0.34
cond_oc_dum_3unemp 0.13
susprindum_oh 0.39
susprindum_coc 0.56
susprindum_pbc 0.35
susprindum_mar 0.18
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.787 1.271 0.537 1.15
less_90d_tr1.lag 1.181 0.847 0.790 1.77
log_dias_treat_imp_sin_na.lag 1.031 0.970 0.789 1.35
comp_bpsc_y3_severe.lag 1.220 0.820 0.896 1.66
policonsumo2.lag 1.234 0.811 0.769 1.98
edad_al_ing_1 1.064 0.940 1.036 1.09
ano_nac_corr 1.064 0.940 1.036 1.09
susinidum_oh 1.312 0.762 0.889 1.94
susinidum_coc 1.119 0.894 0.699 1.79
susinidum_pbc 1.263 0.792 0.814 1.96
susinidum_mar 1.362 0.734 0.915 2.03
psycom_dum_with 1.082 0.924 0.967 1.21
psycom_dum_study 1.077 0.929 0.935 1.24
freq_cons_dum_5day 0.935 1.070 0.835 1.05
cond_oc_dum_2inact 0.930 1.076 0.799 1.08
cond_oc_dum_3unemp 1.096 0.912 0.972 1.24
susprindum_oh 0.852 1.174 0.591 1.23
susprindum_coc 0.893 1.120 0.612 1.30
susprindum_pbc 0.839 1.192 0.582 1.21
susprindum_mar 0.765 1.307 0.517 1.13
Concordance= 0.552 (se = 0.008 )
Likelihood ratio test= 45.3 on 20 df, p=0.001
Wald test = 52 on 20 df, p=1e-04
Score (logrank) test = 44.3 on 20 df, p=0.001, Robust = 50 p=2e-04
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.412 0.552 0.624 0.638 0.705 1.289
[1] "Modality: GP residential\nTime interval: (30,50]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 756, number of events= 756
(698 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.070705 1.073265 0.311271 0.180516 0.39
less_90d_tr1.lag -0.260209 0.770890 0.245527 0.187185 -1.39
log_dias_treat_imp_sin_na.lag -0.055768 0.945758 0.100114 0.081450 -0.68
comp_bpsc_y3_severe.lag 0.785267 2.192992 0.372942 0.268580 2.92
policonsumo2.lag -0.642543 0.525953 0.355505 0.222510 -2.89
edad_al_ing_1 0.081089 1.084468 0.020322 0.018749 4.33
ano_nac_corr 0.084354 1.088014 0.020160 0.018604 4.53
susinidum_oh -0.029896 0.970547 0.247800 0.318855 -0.09
susinidum_coc 0.388930 1.475401 0.325541 0.367282 1.06
susinidum_pbc 0.104636 1.110306 0.265472 0.327583 0.32
susinidum_mar -0.000313 0.999687 0.247905 0.317410 0.00
psycom_dum_with -0.113781 0.892453 0.084083 0.083715 -1.36
psycom_dum_study -0.126070 0.881553 0.117167 0.114303 -1.10
freq_cons_dum_5day 0.007669 1.007698 0.086974 0.081765 0.09
cond_oc_dum_2inact 0.155831 1.168629 0.138600 0.144992 1.07
cond_oc_dum_3unemp 0.239732 1.270908 0.114408 0.107619 2.23
susprindum_oh 0.168151 1.183116 0.337103 0.282967 0.59
susprindum_coc -0.137028 0.871946 0.344384 0.295404 -0.46
susprindum_pbc -0.012679 0.987401 0.326303 0.271486 -0.05
susprindum_mar 0.122759 1.130612 0.377851 0.322122 0.38
Pr(>|z|)
tr_outcome.lag 0.6953
less_90d_tr1.lag 0.1645
log_dias_treat_imp_sin_na.lag 0.4935
comp_bpsc_y3_severe.lag 0.0035 **
policonsumo2.lag 0.0039 **
edad_al_ing_1 1.5e-05 ***
ano_nac_corr 5.8e-06 ***
susinidum_oh 0.9253
susinidum_coc 0.2896
susinidum_pbc 0.7494
susinidum_mar 0.9992
psycom_dum_with 0.1741
psycom_dum_study 0.2700
freq_cons_dum_5day 0.9253
cond_oc_dum_2inact 0.2825
cond_oc_dum_3unemp 0.0259 *
susprindum_oh 0.5523
susprindum_coc 0.6427
susprindum_pbc 0.9628
susprindum_mar 0.7031
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.073 0.932 0.753 1.529
less_90d_tr1.lag 0.771 1.297 0.534 1.113
log_dias_treat_imp_sin_na.lag 0.946 1.057 0.806 1.109
comp_bpsc_y3_severe.lag 2.193 0.456 1.295 3.712
policonsumo2.lag 0.526 1.901 0.340 0.813
edad_al_ing_1 1.084 0.922 1.045 1.125
ano_nac_corr 1.088 0.919 1.049 1.128
susinidum_oh 0.971 1.030 0.520 1.813
susinidum_coc 1.475 0.678 0.718 3.031
susinidum_pbc 1.110 0.901 0.584 2.110
susinidum_mar 1.000 1.000 0.537 1.862
psycom_dum_with 0.892 1.121 0.757 1.052
psycom_dum_study 0.882 1.134 0.705 1.103
freq_cons_dum_5day 1.008 0.992 0.858 1.183
cond_oc_dum_2inact 1.169 0.856 0.880 1.553
cond_oc_dum_3unemp 1.271 0.787 1.029 1.569
susprindum_oh 1.183 0.845 0.679 2.060
susprindum_coc 0.872 1.147 0.489 1.556
susprindum_pbc 0.987 1.013 0.580 1.681
susprindum_mar 1.131 0.884 0.601 2.126
Concordance= 0.567 (se = 0.012 )
Likelihood ratio test= 35.4 on 20 df, p=0.02
Wald test = 54.3 on 20 df, p=5e-05
Score (logrank) test = 35.7 on 20 df, p=0.02, Robust = 43.4 p=0.002
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.356 0.790 0.915 0.932 1.057 2.345
[1] "Modality: WO intensive ambulatory\nTime interval: (30,50]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 272, number of events= 272
(247 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.5814 0.5591 0.6865 0.4366 -1.33
less_90d_tr1.lag 1.2589 3.5215 0.4292 0.3305 3.81
log_dias_treat_imp_sin_na.lag 0.1274 1.1358 0.1779 0.0937 1.36
comp_bpsc_y3_severe.lag -0.8336 0.4345 0.6234 0.4921 -1.69
policonsumo2.lag 0.1199 1.1274 0.6456 0.5892 0.20
edad_al_ing_1 0.0296 1.0301 0.0335 0.0305 0.97
ano_nac_corr 0.0270 1.0274 0.0327 0.0309 0.87
susinidum_oh -0.6095 0.5436 0.4208 0.4329 -1.41
susinidum_coc -0.4464 0.6399 0.5109 0.5645 -0.79
susinidum_pbc -0.3263 0.7216 0.4432 0.4572 -0.71
susinidum_mar -0.3963 0.6728 0.4257 0.4399 -0.90
psycom_dum_with 0.1430 1.1537 0.1442 0.1415 1.01
psycom_dum_study -0.2141 0.8073 0.2120 0.1933 -1.11
freq_cons_dum_5day 0.0414 1.0423 0.1349 0.1394 0.30
cond_oc_dum_2inact 0.0616 1.0635 0.1674 0.1655 0.37
cond_oc_dum_3unemp 0.1078 1.1138 0.1722 0.1674 0.64
susprindum_oh -0.4157 0.6599 0.4391 0.2908 -1.43
susprindum_coc -0.1702 0.8435 0.4485 0.2815 -0.60
susprindum_pbc -0.3561 0.7004 0.4249 0.2678 -1.33
susprindum_mar -0.1563 0.8553 0.4773 0.3353 -0.47
Pr(>|z|)
tr_outcome.lag 0.18298
less_90d_tr1.lag 0.00014 ***
log_dias_treat_imp_sin_na.lag 0.17413
comp_bpsc_y3_severe.lag 0.09030 .
policonsumo2.lag 0.83876
edad_al_ing_1 0.33056
ano_nac_corr 0.38200
susinidum_oh 0.15914
susinidum_coc 0.42909
susinidum_pbc 0.47550
susinidum_mar 0.36775
psycom_dum_with 0.31218
psycom_dum_study 0.26805
freq_cons_dum_5day 0.76662
cond_oc_dum_2inact 0.70979
cond_oc_dum_3unemp 0.51948
susprindum_oh 0.15284
susprindum_coc 0.54539
susprindum_pbc 0.18369
susprindum_mar 0.64124
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.559 1.788 0.238 1.32
less_90d_tr1.lag 3.522 0.284 1.842 6.73
log_dias_treat_imp_sin_na.lag 1.136 0.880 0.945 1.36
comp_bpsc_y3_severe.lag 0.434 2.302 0.166 1.14
policonsumo2.lag 1.127 0.887 0.355 3.58
edad_al_ing_1 1.030 0.971 0.970 1.09
ano_nac_corr 1.027 0.973 0.967 1.09
susinidum_oh 0.544 1.839 0.233 1.27
susinidum_coc 0.640 1.563 0.212 1.93
susinidum_pbc 0.722 1.386 0.295 1.77
susinidum_mar 0.673 1.486 0.284 1.59
psycom_dum_with 1.154 0.867 0.874 1.52
psycom_dum_study 0.807 1.239 0.553 1.18
freq_cons_dum_5day 1.042 0.959 0.793 1.37
cond_oc_dum_2inact 1.064 0.940 0.769 1.47
cond_oc_dum_3unemp 1.114 0.898 0.802 1.55
susprindum_oh 0.660 1.515 0.373 1.17
susprindum_coc 0.843 1.186 0.486 1.46
susprindum_pbc 0.700 1.428 0.414 1.18
susprindum_mar 0.855 1.169 0.443 1.65
Concordance= 0.564 (se = 0.018 )
Likelihood ratio test= 23.1 on 20 df, p=0.3
Wald test = 48.5 on 20 df, p=4e-04
Score (logrank) test = 21.2 on 20 df, p=0.4, Robust = 29.2 p=0.08
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.01 1.69 2.07 2.19 2.44 8.87
[1] "Modality: WO residential\nTime interval: (50,70]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 179, number of events= 179
(161 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.11718 1.12432 0.61305 0.54534 0.21
less_90d_tr1.lag 1.02457 2.78591 1.06311 0.67874 1.51
log_dias_treat_imp_sin_na.lag 0.11300 1.11963 0.72792 0.40985 0.28
comp_bpsc_y3_severe.lag 0.42784 1.53394 0.84312 0.62478 0.68
policonsumo2.lag 0.64360 1.90332 0.56869 0.31440 2.05
edad_al_ing_1 0.03070 1.03118 0.05649 0.05622 0.55
ano_nac_corr 0.04493 1.04596 0.05520 0.05505 0.82
susinidum_oh -0.87341 0.41752 0.63993 0.44994 -1.94
susinidum_coc -0.55173 0.57595 0.72746 0.52218 -1.06
susinidum_pbc -0.76520 0.46524 0.65960 0.46848 -1.63
susinidum_mar -0.42396 0.65445 0.63809 0.47701 -0.89
psycom_dum_with -0.57471 0.56287 0.22561 0.21753 -2.64
psycom_dum_study -0.65671 0.51856 0.26267 0.27317 -2.40
freq_cons_dum_5day 0.01414 1.01425 0.20775 0.20475 0.07
cond_oc_dum_2inact -0.12188 0.88525 0.30041 0.23571 -0.52
cond_oc_dum_3unemp -0.01956 0.98063 0.30449 0.23485 -0.08
susprindum_oh 0.05359 1.05505 0.52266 0.41524 0.13
susprindum_coc 0.24811 1.28159 0.53273 0.42393 0.59
susprindum_pbc 0.05261 1.05402 0.48620 0.39046 0.13
susprindum_mar -0.00979 0.99026 0.77683 0.59414 -0.02
Pr(>|z|)
tr_outcome.lag 0.8299
less_90d_tr1.lag 0.1312
log_dias_treat_imp_sin_na.lag 0.7828
comp_bpsc_y3_severe.lag 0.4935
policonsumo2.lag 0.0406 *
edad_al_ing_1 0.5850
ano_nac_corr 0.4143
susinidum_oh 0.0522 .
susinidum_coc 0.2907
susinidum_pbc 0.1024
susinidum_mar 0.3741
psycom_dum_with 0.0082 **
psycom_dum_study 0.0162 *
freq_cons_dum_5day 0.9449
cond_oc_dum_2inact 0.6051
cond_oc_dum_3unemp 0.9336
susprindum_oh 0.8973
susprindum_coc 0.5584
susprindum_pbc 0.8928
susprindum_mar 0.9869
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.124 0.889 0.386 3.274
less_90d_tr1.lag 2.786 0.359 0.737 10.537
log_dias_treat_imp_sin_na.lag 1.120 0.893 0.501 2.500
comp_bpsc_y3_severe.lag 1.534 0.652 0.451 5.219
policonsumo2.lag 1.903 0.525 1.028 3.525
edad_al_ing_1 1.031 0.970 0.924 1.151
ano_nac_corr 1.046 0.956 0.939 1.165
susinidum_oh 0.418 2.395 0.173 1.008
susinidum_coc 0.576 1.736 0.207 1.603
susinidum_pbc 0.465 2.149 0.186 1.165
susinidum_mar 0.654 1.528 0.257 1.667
psycom_dum_with 0.563 1.777 0.367 0.862
psycom_dum_study 0.519 1.928 0.304 0.886
freq_cons_dum_5day 1.014 0.986 0.679 1.515
cond_oc_dum_2inact 0.885 1.130 0.558 1.405
cond_oc_dum_3unemp 0.981 1.020 0.619 1.554
susprindum_oh 1.055 0.948 0.468 2.381
susprindum_coc 1.282 0.780 0.558 2.942
susprindum_pbc 1.054 0.949 0.490 2.266
susprindum_mar 0.990 1.010 0.309 3.173
Concordance= 0.606 (se = 0.024 )
Likelihood ratio test= 31.8 on 20 df, p=0.05
Wald test = 53.7 on 20 df, p=6e-05
Score (logrank) test = 29.4 on 20 df, p=0.08, Robust = 34 p=0.03
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.131 0.283 0.343 0.426 0.481 2.311
[1] "Modality: GP intensive ambulatory\nTime interval: (50,70]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 952, number of events= 952
(882 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.14373 1.15457 0.40208 0.23848 0.60
less_90d_tr1.lag 0.04075 1.04160 0.26896 0.18882 0.22
log_dias_treat_imp_sin_na.lag -0.06494 0.93712 0.18076 0.14215 -0.46
comp_bpsc_y3_severe.lag 0.24369 1.27594 0.22375 0.15107 1.61
policonsumo2.lag -0.08624 0.91737 0.37665 0.22091 -0.39
edad_al_ing_1 0.09182 1.09616 0.02331 0.02261 4.06
ano_nac_corr 0.08735 1.09128 0.02300 0.02202 3.97
susinidum_oh -0.18020 0.83511 0.18632 0.18875 -0.95
susinidum_coc -0.00580 0.99422 0.23382 0.25265 -0.02
susinidum_pbc 0.09401 1.09857 0.22961 0.22434 0.42
susinidum_mar -0.06294 0.93900 0.19177 0.19659 -0.32
psycom_dum_with 0.11019 1.11649 0.07414 0.07566 1.46
psycom_dum_study 0.07103 1.07361 0.10012 0.09529 0.75
freq_cons_dum_5day 0.03120 1.03170 0.06792 0.06685 0.47
cond_oc_dum_2inact 0.08752 1.09146 0.10398 0.10352 0.85
cond_oc_dum_3unemp -0.00759 0.99244 0.07182 0.07088 -0.11
susprindum_oh -0.33860 0.71277 0.30834 0.26303 -1.29
susprindum_coc -0.49853 0.60742 0.31331 0.26966 -1.85
susprindum_pbc -0.35718 0.69965 0.30406 0.26011 -1.37
susprindum_mar -0.37937 0.68429 0.32550 0.29400 -1.29
Pr(>|z|)
tr_outcome.lag 0.547
less_90d_tr1.lag 0.829
log_dias_treat_imp_sin_na.lag 0.648
comp_bpsc_y3_severe.lag 0.107
policonsumo2.lag 0.696
edad_al_ing_1 4.9e-05 ***
ano_nac_corr 7.3e-05 ***
susinidum_oh 0.340
susinidum_coc 0.982
susinidum_pbc 0.675
susinidum_mar 0.749
psycom_dum_with 0.145
psycom_dum_study 0.456
freq_cons_dum_5day 0.641
cond_oc_dum_2inact 0.398
cond_oc_dum_3unemp 0.915
susprindum_oh 0.198
susprindum_coc 0.064 .
susprindum_pbc 0.170
susprindum_mar 0.197
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.155 0.866 0.723 1.84
less_90d_tr1.lag 1.042 0.960 0.719 1.51
log_dias_treat_imp_sin_na.lag 0.937 1.067 0.709 1.24
comp_bpsc_y3_severe.lag 1.276 0.784 0.949 1.72
policonsumo2.lag 0.917 1.090 0.595 1.41
edad_al_ing_1 1.096 0.912 1.049 1.15
ano_nac_corr 1.091 0.916 1.045 1.14
susinidum_oh 0.835 1.197 0.577 1.21
susinidum_coc 0.994 1.006 0.606 1.63
susinidum_pbc 1.099 0.910 0.708 1.71
susinidum_mar 0.939 1.065 0.639 1.38
psycom_dum_with 1.116 0.896 0.963 1.29
psycom_dum_study 1.074 0.931 0.891 1.29
freq_cons_dum_5day 1.032 0.969 0.905 1.18
cond_oc_dum_2inact 1.091 0.916 0.891 1.34
cond_oc_dum_3unemp 0.992 1.008 0.864 1.14
susprindum_oh 0.713 1.403 0.426 1.19
susprindum_coc 0.607 1.646 0.358 1.03
susprindum_pbc 0.700 1.429 0.420 1.16
susprindum_mar 0.684 1.461 0.385 1.22
Concordance= 0.563 (se = 0.01 )
Likelihood ratio test= 35.3 on 20 df, p=0.02
Wald test = 43.2 on 20 df, p=0.002
Score (logrank) test = 35 on 20 df, p=0.02, Robust = 41.4 p=0.003
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.592 0.927 1.061 1.091 1.220 2.472
[1] "Modality: basic ambulatory\nTime interval: (50,70]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 975, number of events= 975
(912 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.0925 0.9117 0.3173 0.2442 -0.38
less_90d_tr1.lag 0.0706 1.0731 0.2769 0.1839 0.38
log_dias_treat_imp_sin_na.lag 0.0404 1.0412 0.1784 0.1286 0.31
comp_bpsc_y3_severe.lag 0.4059 1.5007 0.2303 0.1791 2.27
policonsumo2.lag 0.1346 1.1441 0.2812 0.2287 0.59
edad_al_ing_1 0.1269 1.1353 0.0210 0.0197 6.45
ano_nac_corr 0.1320 1.1412 0.0210 0.0198 6.67
susinidum_oh -0.0679 0.9344 0.1746 0.2017 -0.34
susinidum_coc -0.3007 0.7403 0.2407 0.2727 -1.10
susinidum_pbc -0.1127 0.8935 0.2201 0.2293 -0.49
susinidum_mar -0.0690 0.9334 0.1813 0.2049 -0.34
psycom_dum_with -0.1255 0.8820 0.0738 0.0725 -1.73
psycom_dum_study -0.0329 0.9677 0.0939 0.0917 -0.36
freq_cons_dum_5day 0.0826 1.0861 0.0754 0.0753 1.10
cond_oc_dum_2inact 0.1227 1.1306 0.1038 0.1018 1.21
cond_oc_dum_3unemp 0.0404 1.0412 0.0784 0.0754 0.54
susprindum_oh 0.2790 1.3218 0.2321 0.2324 1.20
susprindum_coc 0.2343 1.2640 0.2361 0.2345 1.00
susprindum_pbc 0.1675 1.1824 0.2326 0.2319 0.72
susprindum_mar 0.0313 1.0318 0.2554 0.2589 0.12
Pr(>|z|)
tr_outcome.lag 0.705
less_90d_tr1.lag 0.701
log_dias_treat_imp_sin_na.lag 0.753
comp_bpsc_y3_severe.lag 0.023 *
policonsumo2.lag 0.556
edad_al_ing_1 1.1e-10 ***
ano_nac_corr 2.5e-11 ***
susinidum_oh 0.737
susinidum_coc 0.270
susinidum_pbc 0.623
susinidum_mar 0.736
psycom_dum_with 0.083 .
psycom_dum_study 0.720
freq_cons_dum_5day 0.273
cond_oc_dum_2inact 0.228
cond_oc_dum_3unemp 0.592
susprindum_oh 0.230
susprindum_coc 0.318
susprindum_pbc 0.470
susprindum_mar 0.904
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.912 1.097 0.565 1.47
less_90d_tr1.lag 1.073 0.932 0.748 1.54
log_dias_treat_imp_sin_na.lag 1.041 0.960 0.809 1.34
comp_bpsc_y3_severe.lag 1.501 0.666 1.056 2.13
policonsumo2.lag 1.144 0.874 0.731 1.79
edad_al_ing_1 1.135 0.881 1.092 1.18
ano_nac_corr 1.141 0.876 1.098 1.19
susinidum_oh 0.934 1.070 0.629 1.39
susinidum_coc 0.740 1.351 0.434 1.26
susinidum_pbc 0.893 1.119 0.570 1.40
susinidum_mar 0.933 1.071 0.625 1.39
psycom_dum_with 0.882 1.134 0.765 1.02
psycom_dum_study 0.968 1.033 0.808 1.16
freq_cons_dum_5day 1.086 0.921 0.937 1.26
cond_oc_dum_2inact 1.131 0.884 0.926 1.38
cond_oc_dum_3unemp 1.041 0.960 0.898 1.21
susprindum_oh 1.322 0.757 0.838 2.08
susprindum_coc 1.264 0.791 0.798 2.00
susprindum_pbc 1.182 0.846 0.750 1.86
susprindum_mar 1.032 0.969 0.621 1.71
Concordance= 0.577 (se = 0.01 )
Likelihood ratio test= 64.2 on 20 df, p=2e-06
Wald test = 66.3 on 20 df, p=7e-07
Score (logrank) test = 63.6 on 20 df, p=2e-06, Robust = 69.5 p=2e-07
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.290 0.436 0.539 0.559 0.655 1.455
[1] "Modality: GP residential\nTime interval: (50,70]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 392, number of events= 392
(361 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.98955 0.37174 0.45799 0.31909 -3.10
less_90d_tr1.lag 0.83212 2.29819 0.54609 0.48760 1.71
log_dias_treat_imp_sin_na.lag 0.01531 1.01542 0.43790 0.46770 0.03
comp_bpsc_y3_severe.lag -0.20089 0.81800 0.49058 0.31748 -0.63
policonsumo2.lag 0.18496 1.20317 0.60446 0.43474 0.43
edad_al_ing_1 0.09172 1.09606 0.03762 0.03716 2.47
ano_nac_corr 0.10229 1.10770 0.03758 0.03768 2.71
susinidum_oh 0.29856 1.34792 0.28694 0.25281 1.18
susinidum_coc 0.23353 1.26305 0.37920 0.32510 0.72
susinidum_pbc 0.11437 1.12116 0.31636 0.29190 0.39
susinidum_mar 0.05158 1.05293 0.28712 0.25138 0.21
psycom_dum_with -0.20375 0.81567 0.12050 0.11684 -1.74
psycom_dum_study -0.18978 0.82714 0.17528 0.16004 -1.19
freq_cons_dum_5day -0.10802 0.89761 0.11907 0.12143 -0.89
cond_oc_dum_2inact 0.00394 1.00395 0.20053 0.19286 0.02
cond_oc_dum_3unemp -0.10783 0.89778 0.15618 0.15639 -0.69
susprindum_oh -1.69453 0.18369 0.73308 0.20093 -8.43
susprindum_coc -2.05429 0.12818 0.74059 0.22332 -9.20
susprindum_pbc -1.88226 0.15225 0.72721 0.15822 -11.90
susprindum_mar -1.96932 0.13955 0.80891 0.42217 -4.66
Pr(>|z|)
tr_outcome.lag 0.0019 **
less_90d_tr1.lag 0.0879 .
log_dias_treat_imp_sin_na.lag 0.9739
comp_bpsc_y3_severe.lag 0.5269
policonsumo2.lag 0.6705
edad_al_ing_1 0.0136 *
ano_nac_corr 0.0066 **
susinidum_oh 0.2376
susinidum_coc 0.4725
susinidum_pbc 0.6952
susinidum_mar 0.8374
psycom_dum_with 0.0812 .
psycom_dum_study 0.2357
freq_cons_dum_5day 0.3737
cond_oc_dum_2inact 0.9837
cond_oc_dum_3unemp 0.4905
susprindum_oh < 2e-16 ***
susprindum_coc < 2e-16 ***
susprindum_pbc < 2e-16 ***
susprindum_mar 3.1e-06 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.372 2.690 0.1989 0.695
less_90d_tr1.lag 2.298 0.435 0.8838 5.976
log_dias_treat_imp_sin_na.lag 1.015 0.985 0.4060 2.540
comp_bpsc_y3_severe.lag 0.818 1.222 0.4391 1.524
policonsumo2.lag 1.203 0.831 0.5132 2.821
edad_al_ing_1 1.096 0.912 1.0191 1.179
ano_nac_corr 1.108 0.903 1.0288 1.193
susinidum_oh 1.348 0.742 0.8212 2.212
susinidum_coc 1.263 0.792 0.6679 2.389
susinidum_pbc 1.121 0.892 0.6327 1.987
susinidum_mar 1.053 0.950 0.6433 1.723
psycom_dum_with 0.816 1.226 0.6487 1.026
psycom_dum_study 0.827 1.209 0.6044 1.132
freq_cons_dum_5day 0.898 1.114 0.7075 1.139
cond_oc_dum_2inact 1.004 0.996 0.6879 1.465
cond_oc_dum_3unemp 0.898 1.114 0.6608 1.220
susprindum_oh 0.184 5.444 0.1239 0.272
susprindum_coc 0.128 7.801 0.0827 0.199
susprindum_pbc 0.152 6.568 0.1117 0.208
susprindum_mar 0.140 7.166 0.0610 0.319
Concordance= 0.575 (se = 0.017 )
Likelihood ratio test= 32.7 on 20 df, p=0.04
Wald test = 228 on 20 df, p=<2e-16
Score (logrank) test = 34.1 on 20 df, p=0.03, Robust = 31 p=0.06
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.64 7.43 8.78 9.27 10.39 28.37
[1] "Modality: WO intensive ambulatory\nTime interval: (50,70]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 147, number of events= 147
(135 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 2.24572 9.44720 0.97612 0.84362 2.66
less_90d_tr1.lag -1.48459 0.22660 0.79164 0.59000 -2.52
log_dias_treat_imp_sin_na.lag -1.00834 0.36482 0.56973 0.38515 -2.62
comp_bpsc_y3_severe.lag 0.80894 2.24553 0.83762 0.33900 2.39
policonsumo2.lag -1.70058 0.18258 1.35294 0.91105 -1.87
edad_al_ing_1 0.12532 1.13351 0.05847 0.05535 2.26
ano_nac_corr 0.11522 1.12212 0.05606 0.05292 2.18
susinidum_oh -0.39560 0.67327 0.67106 0.38664 -1.02
susinidum_coc 0.06768 1.07002 0.86967 0.72300 0.09
susinidum_pbc -0.74257 0.47589 0.71302 0.48039 -1.55
susinidum_mar -0.35549 0.70083 0.68367 0.43209 -0.82
psycom_dum_with 0.00121 1.00121 0.22120 0.23127 0.01
psycom_dum_study 0.25008 1.28413 0.31634 0.32202 0.78
freq_cons_dum_5day -0.07136 0.93113 0.19453 0.18559 -0.38
cond_oc_dum_2inact 0.13726 1.14713 0.24366 0.26884 0.51
cond_oc_dum_3unemp -0.03216 0.96835 0.25678 0.28903 -0.11
susprindum_oh -0.51243 0.59904 0.78028 0.37638 -1.36
susprindum_coc -0.79785 0.45029 0.81330 0.53602 -1.49
susprindum_pbc -0.39730 0.67213 0.77006 0.40286 -0.99
susprindum_mar -1.05249 0.34907 0.81911 0.46387 -2.27
Pr(>|z|)
tr_outcome.lag 0.0078 **
less_90d_tr1.lag 0.0119 *
log_dias_treat_imp_sin_na.lag 0.0088 **
comp_bpsc_y3_severe.lag 0.0170 *
policonsumo2.lag 0.0620 .
edad_al_ing_1 0.0236 *
ano_nac_corr 0.0295 *
susinidum_oh 0.3062
susinidum_coc 0.9254
susinidum_pbc 0.1222
susinidum_mar 0.4107
psycom_dum_with 0.9958
psycom_dum_study 0.4374
freq_cons_dum_5day 0.7006
cond_oc_dum_2inact 0.6097
cond_oc_dum_3unemp 0.9114
susprindum_oh 0.1734
susprindum_coc 0.1366
susprindum_pbc 0.3240
susprindum_mar 0.0233 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 9.447 0.106 1.8081 49.362
less_90d_tr1.lag 0.227 4.413 0.0713 0.720
log_dias_treat_imp_sin_na.lag 0.365 2.741 0.1715 0.776
comp_bpsc_y3_severe.lag 2.246 0.445 1.1555 4.364
policonsumo2.lag 0.183 5.477 0.0306 1.089
edad_al_ing_1 1.134 0.882 1.0170 1.263
ano_nac_corr 1.122 0.891 1.0116 1.245
susinidum_oh 0.673 1.485 0.3156 1.436
susinidum_coc 1.070 0.935 0.2594 4.414
susinidum_pbc 0.476 2.101 0.1856 1.220
susinidum_mar 0.701 1.427 0.3005 1.635
psycom_dum_with 1.001 0.999 0.6363 1.575
psycom_dum_study 1.284 0.779 0.6831 2.414
freq_cons_dum_5day 0.931 1.074 0.6472 1.340
cond_oc_dum_2inact 1.147 0.872 0.6773 1.943
cond_oc_dum_3unemp 0.968 1.033 0.5496 1.706
susprindum_oh 0.599 1.669 0.2865 1.253
susprindum_coc 0.450 2.221 0.1575 1.288
susprindum_pbc 0.672 1.488 0.3052 1.480
susprindum_mar 0.349 2.865 0.1406 0.866
Concordance= 0.614 (se = 0.025 )
Likelihood ratio test= 29.4 on 20 df, p=0.08
Wald test = 77.1 on 20 df, p=1e-08
Score (logrank) test = 25.8 on 20 df, p=0.2, Robust = 28.5 p=0.1
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.0 2.0 2.6 3.2 3.3 42.1
[1] "Modality: WO residential\nTime interval: (70,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 140, number of events= 140
(114 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.2111 1.2351 0.5133 0.3758 0.56
less_90d_tr1.lag 0.1271 1.1356 0.4777 0.3623 0.35
log_dias_treat_imp_sin_na.lag 0.0950 1.0997 0.2702 0.1604 0.59
comp_bpsc_y3_severe.lag -0.0265 0.9738 0.5683 0.4003 -0.07
policonsumo2.lag 0.4960 1.6421 0.7126 0.5910 0.84
edad_al_ing_1 0.7531 2.1235 0.1295 0.1259 5.98
ano_nac_corr 0.7581 2.1342 0.1267 0.1213 6.25
susinidum_oh -0.1409 0.8686 0.7396 0.6574 -0.21
susinidum_coc 0.2827 1.3267 0.9117 0.8001 0.35
susinidum_pbc 0.0692 1.0716 0.7778 0.7062 0.10
susinidum_mar -0.4544 0.6348 0.7461 0.6705 -0.68
psycom_dum_with 0.2074 1.2304 0.2479 0.2219 0.93
psycom_dum_study 0.3116 1.3656 0.3006 0.2733 1.14
freq_cons_dum_5day -0.1393 0.8700 0.2268 0.2203 -0.63
cond_oc_dum_2inact 0.3615 1.4355 0.3814 0.2760 1.31
cond_oc_dum_3unemp 0.3359 1.3992 0.3732 0.2670 1.26
susprindum_oh -0.8623 0.4222 0.6834 0.5388 -1.60
susprindum_coc -0.7442 0.4751 0.6847 0.5885 -1.26
susprindum_pbc -0.8763 0.4163 0.6689 0.5671 -1.55
susprindum_mar -0.1224 0.8848 0.8132 0.7029 -0.17
Pr(>|z|)
tr_outcome.lag 0.57
less_90d_tr1.lag 0.73
log_dias_treat_imp_sin_na.lag 0.55
comp_bpsc_y3_severe.lag 0.95
policonsumo2.lag 0.40
edad_al_ing_1 2.2e-09 ***
ano_nac_corr 4.1e-10 ***
susinidum_oh 0.83
susinidum_coc 0.72
susinidum_pbc 0.92
susinidum_mar 0.50
psycom_dum_with 0.35
psycom_dum_study 0.25
freq_cons_dum_5day 0.53
cond_oc_dum_2inact 0.19
cond_oc_dum_3unemp 0.21
susprindum_oh 0.11
susprindum_coc 0.21
susprindum_pbc 0.12
susprindum_mar 0.86
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.235 0.810 0.591 2.58
less_90d_tr1.lag 1.136 0.881 0.558 2.31
log_dias_treat_imp_sin_na.lag 1.100 0.909 0.803 1.51
comp_bpsc_y3_severe.lag 0.974 1.027 0.444 2.13
policonsumo2.lag 1.642 0.609 0.516 5.23
edad_al_ing_1 2.123 0.471 1.659 2.72
ano_nac_corr 2.134 0.469 1.683 2.71
susinidum_oh 0.869 1.151 0.239 3.15
susinidum_coc 1.327 0.754 0.277 6.37
susinidum_pbc 1.072 0.933 0.268 4.28
susinidum_mar 0.635 1.575 0.171 2.36
psycom_dum_with 1.230 0.813 0.797 1.90
psycom_dum_study 1.366 0.732 0.799 2.33
freq_cons_dum_5day 0.870 1.149 0.565 1.34
cond_oc_dum_2inact 1.436 0.697 0.836 2.47
cond_oc_dum_3unemp 1.399 0.715 0.829 2.36
susprindum_oh 0.422 2.368 0.147 1.21
susprindum_coc 0.475 2.105 0.150 1.51
susprindum_pbc 0.416 2.402 0.137 1.27
susprindum_mar 0.885 1.130 0.223 3.51
Concordance= 0.692 (se = 0.021 )
Likelihood ratio test= 60.2 on 20 df, p=7e-06
Wald test = 101 on 20 df, p=8e-13
Score (logrank) test = 66.8 on 20 df, p=6e-07, Robust = 63.5 p=2e-06
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.11 0.47 1.03 1.17 1.55 4.14
[1] "Modality: GP intensive ambulatory\nTime interval: (70,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 747, number of events= 747
(659 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.10065 0.90425 0.27513 0.19992 -0.50
less_90d_tr1.lag 0.22444 1.25162 0.28593 0.20305 1.11
log_dias_treat_imp_sin_na.lag 0.08647 1.09032 0.20200 0.14347 0.60
comp_bpsc_y3_severe.lag 0.08014 1.08344 0.21268 0.15931 0.50
policonsumo2.lag 0.28875 1.33476 0.26990 0.18411 1.57
edad_al_ing_1 0.53755 1.71182 0.03877 0.03623 14.84
ano_nac_corr 0.53650 1.71001 0.03879 0.03598 14.91
susinidum_oh -0.08569 0.91788 0.23039 0.22759 -0.38
susinidum_coc 0.01008 1.01013 0.29231 0.28871 0.03
susinidum_pbc 0.00295 1.00295 0.28632 0.27217 0.01
susinidum_mar 0.01616 1.01629 0.23648 0.23409 0.07
psycom_dum_with -0.11452 0.89179 0.08329 0.08055 -1.42
psycom_dum_study -0.01216 0.98792 0.11479 0.10612 -0.11
freq_cons_dum_5day -0.01605 0.98408 0.07879 0.07317 -0.22
cond_oc_dum_2inact 0.07352 1.07629 0.12276 0.11230 0.65
cond_oc_dum_3unemp 0.02951 1.02995 0.08323 0.07784 0.38
susprindum_oh -0.06372 0.93827 0.33425 0.18919 -0.34
susprindum_coc -0.06906 0.93327 0.33636 0.19413 -0.36
susprindum_pbc -0.11980 0.88709 0.32915 0.17929 -0.67
susprindum_mar -0.04056 0.96025 0.35340 0.22979 -0.18
Pr(>|z|)
tr_outcome.lag 0.61
less_90d_tr1.lag 0.27
log_dias_treat_imp_sin_na.lag 0.55
comp_bpsc_y3_severe.lag 0.61
policonsumo2.lag 0.12
edad_al_ing_1 <2e-16 ***
ano_nac_corr <2e-16 ***
susinidum_oh 0.71
susinidum_coc 0.97
susinidum_pbc 0.99
susinidum_mar 0.94
psycom_dum_with 0.16
psycom_dum_study 0.91
freq_cons_dum_5day 0.83
cond_oc_dum_2inact 0.51
cond_oc_dum_3unemp 0.70
susprindum_oh 0.74
susprindum_coc 0.72
susprindum_pbc 0.50
susprindum_mar 0.86
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.904 1.106 0.611 1.34
less_90d_tr1.lag 1.252 0.799 0.841 1.86
log_dias_treat_imp_sin_na.lag 1.090 0.917 0.823 1.44
comp_bpsc_y3_severe.lag 1.083 0.923 0.793 1.48
policonsumo2.lag 1.335 0.749 0.930 1.91
edad_al_ing_1 1.712 0.584 1.594 1.84
ano_nac_corr 1.710 0.585 1.594 1.83
susinidum_oh 0.918 1.089 0.588 1.43
susinidum_coc 1.010 0.990 0.574 1.78
susinidum_pbc 1.003 0.997 0.588 1.71
susinidum_mar 1.016 0.984 0.642 1.61
psycom_dum_with 0.892 1.121 0.762 1.04
psycom_dum_study 0.988 1.012 0.802 1.22
freq_cons_dum_5day 0.984 1.016 0.853 1.14
cond_oc_dum_2inact 1.076 0.929 0.864 1.34
cond_oc_dum_3unemp 1.030 0.971 0.884 1.20
susprindum_oh 0.938 1.066 0.648 1.36
susprindum_coc 0.933 1.072 0.638 1.37
susprindum_pbc 0.887 1.127 0.624 1.26
susprindum_mar 0.960 1.041 0.612 1.51
Concordance= 0.643 (se = 0.011 )
Likelihood ratio test= 214 on 20 df, p=<2e-16
Wald test = 271 on 20 df, p=<2e-16
Score (logrank) test = 217 on 20 df, p=<2e-16, Robust = 197 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.12 0.48 0.82 0.93 1.24 3.75
[1] "Modality: basic ambulatory\nTime interval: (70,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 699, number of events= 699
(617 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.1665 0.8466 0.2975 0.2632 -0.63
less_90d_tr1.lag -0.4788 0.6195 0.2536 0.2275 -2.10
log_dias_treat_imp_sin_na.lag -0.3879 0.6785 0.1119 0.0973 -3.99
comp_bpsc_y3_severe.lag 0.3991 1.4905 0.2041 0.1726 2.31
policonsumo2.lag -0.0987 0.9060 0.2908 0.2904 -0.34
edad_al_ing_1 0.5149 1.6735 0.0418 0.0407 12.64
ano_nac_corr 0.5131 1.6704 0.0414 0.0406 12.65
susinidum_oh -0.1703 0.8434 0.2424 0.2253 -0.76
susinidum_coc -0.3957 0.6732 0.2808 0.2530 -1.56
susinidum_pbc -0.1599 0.8523 0.2782 0.2542 -0.63
susinidum_mar -0.1832 0.8326 0.2505 0.2308 -0.79
psycom_dum_with 0.0881 1.0921 0.0869 0.0819 1.08
psycom_dum_study -0.0535 0.9479 0.1099 0.0978 -0.55
freq_cons_dum_5day -0.0173 0.9829 0.0880 0.0812 -0.21
cond_oc_dum_2inact 0.1352 1.1448 0.1208 0.1225 1.10
cond_oc_dum_3unemp -0.1115 0.8945 0.0954 0.0841 -1.32
susprindum_oh -0.2414 0.7855 0.3180 0.2378 -1.02
susprindum_coc -0.1284 0.8795 0.3166 0.2318 -0.55
susprindum_pbc -0.1876 0.8290 0.3126 0.2252 -0.83
susprindum_mar -0.0853 0.9183 0.3332 0.2448 -0.35
Pr(>|z|)
tr_outcome.lag 0.527
less_90d_tr1.lag 0.035 *
log_dias_treat_imp_sin_na.lag 6.7e-05 ***
comp_bpsc_y3_severe.lag 0.021 *
policonsumo2.lag 0.734
edad_al_ing_1 < 2e-16 ***
ano_nac_corr < 2e-16 ***
susinidum_oh 0.450
susinidum_coc 0.118
susinidum_pbc 0.529
susinidum_mar 0.427
psycom_dum_with 0.282
psycom_dum_study 0.585
freq_cons_dum_5day 0.832
cond_oc_dum_2inact 0.270
cond_oc_dum_3unemp 0.185
susprindum_oh 0.310
susprindum_coc 0.580
susprindum_pbc 0.405
susprindum_mar 0.728
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.847 1.181 0.505 1.418
less_90d_tr1.lag 0.620 1.614 0.397 0.968
log_dias_treat_imp_sin_na.lag 0.679 1.474 0.561 0.821
comp_bpsc_y3_severe.lag 1.491 0.671 1.063 2.091
policonsumo2.lag 0.906 1.104 0.513 1.601
edad_al_ing_1 1.674 0.598 1.545 1.813
ano_nac_corr 1.670 0.599 1.543 1.809
susinidum_oh 0.843 1.186 0.542 1.312
susinidum_coc 0.673 1.485 0.410 1.105
susinidum_pbc 0.852 1.173 0.518 1.403
susinidum_mar 0.833 1.201 0.530 1.309
psycom_dum_with 1.092 0.916 0.930 1.282
psycom_dum_study 0.948 1.055 0.783 1.148
freq_cons_dum_5day 0.983 1.017 0.838 1.152
cond_oc_dum_2inact 1.145 0.874 0.900 1.455
cond_oc_dum_3unemp 0.895 1.118 0.759 1.055
susprindum_oh 0.785 1.273 0.493 1.252
susprindum_coc 0.880 1.137 0.558 1.385
susprindum_pbc 0.829 1.206 0.533 1.289
susprindum_mar 0.918 1.089 0.568 1.484
Concordance= 0.628 (se = 0.012 )
Likelihood ratio test= 173 on 20 df, p=<2e-16
Wald test = 198 on 20 df, p=<2e-16
Score (logrank) test = 174 on 20 df, p=<2e-16, Robust = 151 p=<2e-16
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.37 1.31 2.03 2.33 3.02 12.32
[1] "Modality: GP residential\nTime interval: (70,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 325, number of events= 325
(273 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag -0.1283 0.8796 0.3809 0.2124 -0.60
less_90d_tr1.lag 0.2810 1.3244 0.3371 0.2433 1.15
log_dias_treat_imp_sin_na.lag -0.0898 0.9141 0.2532 0.1974 -0.45
comp_bpsc_y3_severe.lag 0.3132 1.3678 0.3254 0.1937 1.62
policonsumo2.lag -0.2731 0.7610 0.3553 0.2857 -0.96
edad_al_ing_1 0.6022 1.8261 0.0657 0.0567 10.61
ano_nac_corr 0.6002 1.8225 0.0658 0.0564 10.64
susinidum_oh -0.5709 0.5650 0.3978 0.2116 -2.70
susinidum_coc -0.4843 0.6162 0.4715 0.2666 -1.82
susinidum_pbc -0.6679 0.5128 0.4202 0.2403 -2.78
susinidum_mar -0.7051 0.4941 0.4015 0.2205 -3.20
psycom_dum_with 0.0729 1.0756 0.1290 0.1183 0.62
psycom_dum_study 0.1026 1.1081 0.1852 0.1607 0.64
freq_cons_dum_5day -0.1321 0.8762 0.1262 0.1142 -1.16
cond_oc_dum_2inact -0.3295 0.7193 0.2323 0.1969 -1.67
cond_oc_dum_3unemp -0.3918 0.6758 0.1649 0.1471 -2.66
susprindum_oh 0.1714 1.1869 0.7397 0.3254 0.53
susprindum_coc 0.1972 1.2179 0.7364 0.3188 0.62
susprindum_pbc 0.2918 1.3389 0.7228 0.2736 1.07
susprindum_mar 0.2866 1.3319 0.8015 0.3709 0.77
Pr(>|z|)
tr_outcome.lag 0.5457
less_90d_tr1.lag 0.2483
log_dias_treat_imp_sin_na.lag 0.6492
comp_bpsc_y3_severe.lag 0.1059
policonsumo2.lag 0.3391
edad_al_ing_1 <2e-16 ***
ano_nac_corr <2e-16 ***
susinidum_oh 0.0070 **
susinidum_coc 0.0693 .
susinidum_pbc 0.0054 **
susinidum_mar 0.0014 **
psycom_dum_with 0.5378
psycom_dum_study 0.5231
freq_cons_dum_5day 0.2475
cond_oc_dum_2inact 0.0943 .
cond_oc_dum_3unemp 0.0077 **
susprindum_oh 0.5985
susprindum_coc 0.5363
susprindum_pbc 0.2861
susprindum_mar 0.4397
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 0.880 1.137 0.580 1.334
less_90d_tr1.lag 1.324 0.755 0.822 2.134
log_dias_treat_imp_sin_na.lag 0.914 1.094 0.621 1.346
comp_bpsc_y3_severe.lag 1.368 0.731 0.936 1.999
policonsumo2.lag 0.761 1.314 0.435 1.332
edad_al_ing_1 1.826 0.548 1.634 2.041
ano_nac_corr 1.822 0.549 1.632 2.036
susinidum_oh 0.565 1.770 0.373 0.855
susinidum_coc 0.616 1.623 0.365 1.039
susinidum_pbc 0.513 1.950 0.320 0.821
susinidum_mar 0.494 2.024 0.321 0.761
psycom_dum_with 1.076 0.930 0.853 1.356
psycom_dum_study 1.108 0.902 0.809 1.518
freq_cons_dum_5day 0.876 1.141 0.700 1.096
cond_oc_dum_2inact 0.719 1.390 0.489 1.058
cond_oc_dum_3unemp 0.676 1.480 0.507 0.902
susprindum_oh 1.187 0.843 0.627 2.246
susprindum_coc 1.218 0.821 0.652 2.275
susprindum_pbc 1.339 0.747 0.783 2.289
susprindum_mar 1.332 0.751 0.644 2.755
Concordance= 0.652 (se = 0.015 )
Likelihood ratio test= 96 on 20 df, p=7e-12
Wald test = 144 on 20 df, p=<2e-16
Score (logrank) test = 99.3 on 20 df, p=2e-12, Robust = 100 p=1e-12
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.21 1.13 1.84 2.11 2.79 6.63
[1] "Modality: WO intensive ambulatory\nTime interval: (70,135]"
Call:
coxph(formula = Surv(time.lag, time, event) ~ tr_outcome.lag +
less_90d_tr1.lag + log_dias_treat_imp_sin_na.lag + comp_bpsc_y3_severe.lag +
policonsumo2.lag + edad_al_ing_1 + ano_nac_corr + susinidum_oh +
susinidum_coc + susinidum_pbc + susinidum_mar + psycom_dum_with +
psycom_dum_study + freq_cons_dum_5day + cond_oc_dum_2inact +
cond_oc_dum_3unemp + susprindum_oh + susprindum_coc + susprindum_pbc +
susprindum_mar, data = datacox, cluster = id)
n= 109, number of events= 109
(89 observations deleted due to missingness)
coef exp(coef) se(coef) robust se z
tr_outcome.lag 0.6069 1.8347 0.6968 0.5730 1.06
less_90d_tr1.lag 0.9729 2.6457 0.7169 0.4316 2.25
log_dias_treat_imp_sin_na.lag -0.0350 0.9656 0.2175 0.0989 -0.35
comp_bpsc_y3_severe.lag -0.5660 0.5678 0.6461 0.4651 -1.22
policonsumo2.lag -1.1665 0.3114 0.7091 0.4960 -2.35
edad_al_ing_1 0.7256 2.0660 0.1272 0.1194 6.08
ano_nac_corr 0.7006 2.0149 0.1234 0.1180 5.94
susinidum_oh 0.1262 1.1345 1.1294 0.4389 0.29
susinidum_coc -1.3507 0.2591 1.5526 0.5259 -2.57
susinidum_pbc 0.6761 1.9662 1.1723 0.5557 1.22
susinidum_mar 0.0301 1.0305 1.1478 0.5383 0.06
psycom_dum_with -0.0459 0.9552 0.2525 0.2479 -0.19
psycom_dum_study 0.3213 1.3789 0.3185 0.2964 1.08
freq_cons_dum_5day 0.3997 1.4914 0.2730 0.2428 1.65
cond_oc_dum_2inact 0.3015 1.3519 0.3518 0.3294 0.92
cond_oc_dum_3unemp 0.1458 1.1570 0.3493 0.3020 0.48
susprindum_oh 0.7995 2.2245 0.8200 0.5875 1.36
susprindum_coc 0.6280 1.8739 0.8374 0.5909 1.06
susprindum_pbc 0.7274 2.0697 0.7852 0.5697 1.28
susprindum_mar 1.1757 3.2405 0.9061 0.6070 1.94
Pr(>|z|)
tr_outcome.lag 0.289
less_90d_tr1.lag 0.024 *
log_dias_treat_imp_sin_na.lag 0.723
comp_bpsc_y3_severe.lag 0.224
policonsumo2.lag 0.019 *
edad_al_ing_1 1.2e-09 ***
ano_nac_corr 2.9e-09 ***
susinidum_oh 0.774
susinidum_coc 0.010 *
susinidum_pbc 0.224
susinidum_mar 0.955
psycom_dum_with 0.853
psycom_dum_study 0.278
freq_cons_dum_5day 0.100 .
cond_oc_dum_2inact 0.360
cond_oc_dum_3unemp 0.629
susprindum_oh 0.174
susprindum_coc 0.288
susprindum_pbc 0.202
susprindum_mar 0.053 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
exp(coef) exp(-coef) lower .95 upper .95
tr_outcome.lag 1.835 0.545 0.5968 5.640
less_90d_tr1.lag 2.646 0.378 1.1354 6.165
log_dias_treat_imp_sin_na.lag 0.966 1.036 0.7954 1.172
comp_bpsc_y3_severe.lag 0.568 1.761 0.2282 1.413
policonsumo2.lag 0.311 3.211 0.1178 0.823
edad_al_ing_1 2.066 0.484 1.6348 2.611
ano_nac_corr 2.015 0.496 1.5988 2.539
susinidum_oh 1.135 0.881 0.4800 2.682
susinidum_coc 0.259 3.860 0.0924 0.726
susinidum_pbc 1.966 0.509 0.6616 5.844
susinidum_mar 1.031 0.970 0.3588 2.960
psycom_dum_with 0.955 1.047 0.5876 1.553
psycom_dum_study 1.379 0.725 0.7714 2.465
freq_cons_dum_5day 1.491 0.671 0.9267 2.400
cond_oc_dum_2inact 1.352 0.740 0.7089 2.578
cond_oc_dum_3unemp 1.157 0.864 0.6401 2.091
susprindum_oh 2.225 0.450 0.7033 7.036
susprindum_coc 1.874 0.534 0.5886 5.966
susprindum_pbc 2.070 0.483 0.6777 6.321
susprindum_mar 3.241 0.309 0.9861 10.649
Concordance= 0.692 (se = 0.025 )
Likelihood ratio test= 50.8 on 20 df, p=2e-04
Wald test = 237 on 20 df, p=<2e-16
Score (logrank) test = 48.5 on 20 df, p=4e-04, Robust = 40.6 p=0.004
(Note: the likelihood ratio and score tests assume independence of
observations within a cluster, the Wald and robust score tests do not).
Min. 1st Qu. Median Mean 3rd Qu. Max.
0.01 0.19 0.35 0.62 0.65 9.87 Code
data_mine_miss_proc2_iiw_strata <- NULL # Initialize empty result
for (i in 1:nrow(comb_strata)) {
subset_data <- subset(dplyr::mutate(data_mine_miss_restr_proc2, rn=dplyr::row_number()), subset= tipo_de_plan_2_mod==comb_strata[i,"Var1"] & time_interval3== comb_strata[i,"Var2"])
subset_data2<- cbind.data.frame(subset_data, iiw_strata= iiws_strat[[i]][[1]][[1]])
data_mine_miss_proc2_iiw_strata <- rbind.data.frame(data_mine_miss_proc2_iiw_strata, subset_data2)
}
#:#:#:#:#:#:#:#:#:#:#:#:#
#:#:#:#:#:#:#:#:#:#:#:#:#
#:#:#:#:#:#:#:#:#:#:#:#:#
data_mine_miss_proc2_iiw_strata_alt <- NULL # Initialize empty result
for (i in 1:nrow(comb_strata_alt)) {
subset_data_alt <- subset(dplyr::mutate(data_mine_miss_restr_proc2, rn=dplyr::row_number()), subset= tipo_de_plan_2_mod==comb_strata_alt[i,"Var1"] & time_interval3_alt== comb_strata_alt[i,"Var2"])
subset_data2_alt<- cbind.data.frame(subset_data_alt, iiw_strata= iiws_strat_alt[[i]][[1]][[1]])
data_mine_miss_proc2_iiw_strata_alt <- rbind.data.frame(data_mine_miss_proc2_iiw_strata_alt, subset_data2_alt)
}
#:#:#:#:#:#:#:#:#:#:#:#:#
#:#:#:#:#:#:#:#:#:#:#:#:#
#:#:#:#:#:#:#:#:#:#:#:#:#
data_mine_miss_proc2_iiw_strata_alt_alt <- NULL # Initialize empty result
for (i in 1:nrow(comb_strata_alt_alt)) {
subset_data_alt <- subset(dplyr::mutate(data_mine_miss_restr_proc2, rn=dplyr::row_number()), subset= tipo_de_plan_2_mod==comb_strata_alt_alt[i,"Var1"] & time_interval3_alt== comb_strata_alt_alt[i,"Var2"])
subset_data2_alt_alt<- cbind.data.frame(subset_data_alt, iiw_strata= iiws_strat_alt_alt[[i]][[1]][[1]])
data_mine_miss_proc2_iiw_strata_alt_alt <- rbind.data.frame(data_mine_miss_proc2_iiw_strata_alt_alt, subset_data2_alt_alt)
}
#:#:#:#:#:#:#:#:#:#:#:#:#
#:#:#:#:#:#:#:#:#:#:#:#:#
#:#:#:#:#:#:#:#:#:#:#:#:#
summary(data_mine_miss_proc2_iiw_strata$iiw_strata) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.00e+00 1.00e+00 1.00e+00 9.60e+64 3.00e+00 2.98e+69 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.18 0.77 0.91 0.91 1.04 10.94
summary(data_mine_miss_proc2_iiw_strata_alt$iiw_strata) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.01 0.60 0.86 1.07 1.18 26.75 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.01 0.61 0.87 1.08 1.20 29.96
summary(data_mine_miss_proc2_iiw_strata_alt_alt$iiw_strata) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.0 0.3 0.4 0.7 0.8 42.1 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.0 0.3 0.4 0.7 0.8 42.1
invisible("Más confiable el 2do por LEJOS")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Stabilize weights")
data_mine_miss_proc2_iiw_strata$iiw_strata_st<-data_mine_miss_proc2_iiw_strata$iiw_strata
data_mine_miss_proc2_iiw_strata$iiw_strata_st[data_mine_miss_proc2_iiw_strata$iiw_strata>quantile(data_mine_miss_proc2_iiw_strata$iiw_strata,0.975)] <- quantile(data_mine_miss_proc2_iiw_strata$iiw_strata,0.975)
data_mine_miss_proc2_iiw_strata$iiw_strata_st[data_mine_miss_proc2_iiw_strata$iiw_after_ph<quantile(data_mine_miss_proc2_iiw_strata$iiw_strata,0.025)] <- quantile(data_mine_miss_proc2_iiw_strata$iiw_strata,0.025)
invisible("stabilize alternative weights")
data_mine_miss_proc2_iiw_strata_alt$iiw_strata_st<-data_mine_miss_proc2_iiw_strata_alt$iiw_strata
data_mine_miss_proc2_iiw_strata_alt$iiw_strata_st[data_mine_miss_proc2_iiw_strata_alt$iiw_strata>quantile(data_mine_miss_proc2_iiw_strata_alt$iiw_strata,0.975)] <- quantile(data_mine_miss_proc2_iiw_strata_alt$iiw_strata,0.975)
data_mine_miss_proc2_iiw_strata_alt$iiw_strata_st[data_mine_miss_proc2_iiw_strata_alt$iiw_strata<quantile(data_mine_miss_proc2_iiw_strata_alt$iiw_strata,0.025)] <- quantile(data_mine_miss_proc2_iiw_strata_alt$iiw_strata,0.025)
data_mine_miss_proc2_iiw_strata_alt_alt$iiw_strata_st<-data_mine_miss_proc2_iiw_strata_alt_alt$iiw_strata
data_mine_miss_proc2_iiw_strata_alt_alt$iiw_strata_st[data_mine_miss_proc2_iiw_strata_alt_alt$iiw_strata>quantile(data_mine_miss_proc2_iiw_strata_alt$iiw_strata,0.975)] <- quantile(data_mine_miss_proc2_iiw_strata_alt_alt$iiw_strata,0.975)
data_mine_miss_proc2_iiw_strata_alt_alt$iiw_strata_st[data_mine_miss_proc2_iiw_strata_alt_alt$iiw_strata<quantile(data_mine_miss_proc2_iiw_strata_alt$iiw_strata,0.025)] <- quantile(data_mine_miss_proc2_iiw_strata_alt_alt$iiw_strata,0.025)
summary(data_mine_miss_proc2_iiw_strata$iiw_strata_st) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.00 0.60 1.32 1.73 2.56 5.45 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.00 0.78 1.37 1.88 2.67 5.68
summary(data_mine_miss_proc2_iiw_strata_alt$iiw_strata_st) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.288 0.599 0.863 0.987 1.184 3.018 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.292 0.614 0.873 1.003 1.196 3.121
summary(data_mine_miss_proc2_iiw_strata_alt_alt$iiw_strata_st) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.116 0.116 0.439 0.605 0.761 3.069 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.116 0.116 0.439 0.603 0.761 3.118
rbind.data.frame(
cbind.data.frame(model= "Primary (strata, lag=0)", t(matrix(summary(data_mine_miss_proc2_iiw_strata$iiw_strata_st)))),
cbind.data.frame(model= "Alternative (strata, lag=1)", t(matrix(summary(data_mine_miss_proc2_iiw_strata_alt$iiw_strata_st)))),
cbind.data.frame(model= "Alternative (2nd strata intervals, lag=1)", t(matrix(summary(data_mine_miss_proc2_iiw_strata_alt$iiw_strata_st)))))%>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/iiw_strat_240503.csv"), dec=",", sep="\t")
knitr::kable(., size=10, format="markdown", caption="Weights (stratifying)", col.names= c("Weight",attr(summary(data_mine_miss_proc2_iiw_strata$iiw_strata_st),"names")) )
}| Weight | Min. | 1st Qu. | Median | Mean | 3rd Qu. | Max. |
|---|---|---|---|---|---|---|
| Primary (strata, lag=0) | 0.000 | 0.599 | 1.324 | 1.734 | 2.56 | 5.45 |
| Alternative (strata, lag=1) | 0.288 | 0.599 | 0.863 | 0.987 | 1.18 | 3.02 |
| Alternative (2nd strata intervals, lag=1) | 0.288 | 0.599 | 0.863 | 0.987 | 1.18 | 3.02 |
GEE
Code
#The id= option is where we specify the clusters within which we have repeated observations that may be correlated, and tehe scale.fix=T option is to avoid R's default approach of introducing an overdispersion scale parameter.
#However, caution is advised when employing QIC and its use should not be routine, see Wang et al. (2015). growing use of quasi-likelihood-based information criteria for longitudinal data to select a working correlation structure in a generalized estimating equation framework. [https://onlinelibrary.wiley.com/doi/full/10.1002/sta4.95]
#poisson(link = "log"),
#https://www.thelancet.com/cms/10.1016/S2468-2667(22)00042-1/attachment/eb69ffc4-cb6c-4a47-9524-94e1ba9302b2/mmc1.pdf
#https://www.thelancet.com/cms/10.1016/S2468-2667(22)00201-8/attachment/05a17ae1-1e31-4ffd-a727-78ec915dc03b/mmc1.pdf
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Must be independence structure")
m_strata <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_proc2_iiw_strata, family=binomial, corstr="independence")
summary(m)
Call:
geeglm(formula = tr_outcome ~ policonsumo2 + edad_al_ing_1 +
ano_nac_corr + susinidum_oh + susinidum_coc + susinidum_pbc +
susinidum_mar + psycom_dum_study + psycom_dum_with + freq_cons_dum_5day +
cond_oc_dum_2inact + cond_oc_dum_3unemp + susprindum_coc +
susprindum_pbc + susprindum_mar, family = binomial, data = data_mine_miss_restr_proc2_iiw_after_ph,
id = id, corstr = "independence")
Coefficients:
Estimate Std.err Wald Pr(>|W|)
(Intercept) -43.33863 13.16878 10.83 0.00100 ***
policonsumo2 0.05760 0.03597 2.56 0.10927
edad_al_ing_1 -0.00669 0.00668 1.01 0.31608
ano_nac_corr 0.02251 0.00654 11.83 0.00058 ***
susinidum_oh -0.07388 0.09434 0.61 0.43356
susinidum_coc -0.01507 0.11807 0.02 0.89845
susinidum_pbc 0.14542 0.10781 1.82 0.17737
susinidum_mar 0.04749 0.09658 0.24 0.62290
psycom_dum_study 0.86346 0.04632 347.54 < 2e-16 ***
psycom_dum_with 0.06046 0.03263 3.43 0.06392 .
freq_cons_dum_5day -0.02979 0.03089 0.93 0.33480
cond_oc_dum_2inact -0.25843 0.04290 36.29 1.7e-09 ***
cond_oc_dum_3unemp -0.19282 0.03449 31.25 2.3e-08 ***
susprindum_coc 0.02959 0.04840 0.37 0.54105
susprindum_pbc 0.23707 0.04089 33.61 6.7e-09 ***
susprindum_mar 0.06276 0.07496 0.70 0.40244
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation structure = independence
Estimated Scale Parameters:
Estimate Std.err
(Intercept) 1 0.0162
Number of clusters: 13317 Maximum cluster size: 10 Code
m_strata_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_proc2_iiw_strata, family=binomial, corstr="independence")
rbind.data.frame(cbind.data.frame(model="simplest", broom::tidy(m_strata, exponentiate = T, conf.int=T)),cbind.data.frame(model="full", broom::tidy(m_strata_full, exponentiate = T, conf.int=T))) %>% dplyr::mutate_at(c("estimate","std.error","conf.low","conf.high"), ~sprintf("%1.2f",.))%>% dplyr::mutate_at(c("p.value"), ~sprintf("%1.4f",.)) %>% dplyr::select(-statistic) %>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/gee_strata_240503.csv"), dec=",", sep="\t")
knitr::kable(dplyr::filter(.,term=="policonsumo2"), size=10, format="markdown", caption="GEE Models (Stratifying)")
}| model | term | estimate | std.error | p.value | conf.low | conf.high |
|---|---|---|---|---|---|---|
| simplest | policonsumo2 | 1.06 | 0.04 | 0.1095 | 0.99 | 1.14 |
| full | policonsumo2 | 1.10 | 0.04 | 0.0080 | 1.03 | 1.18 |
Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Call:
geeglm(formula = tr_outcome ~ policonsumo2 + edad_al_ing_1 +
ano_nac_corr + susinidum_oh + susinidum_coc + susinidum_pbc +
susinidum_mar + psycom_dum_study + psycom_dum_with + freq_cons_dum_5day +
cond_oc_dum_2inact + cond_oc_dum_3unemp + susprindum_coc +
susprindum_pbc + susprindum_mar, family = poisson(link = "log"),
data = data_mine_miss_proc2_iiw_strata, weights = iiw_strata_st,
id = id, corstr = "independence")
Coefficients:
Estimate Std.err Wald Pr(>|W|)
(Intercept) -5.66e+00 3.86e+00 2.15 0.14247
policonsumo2 2.83e-02 1.18e-02 5.74 0.01660 *
edad_al_ing_1 -5.53e-03 1.96e-03 7.98 0.00473 **
ano_nac_corr 2.77e-03 1.92e-03 2.09 0.14849
susinidum_oh 3.19e-03 3.21e-02 0.01 0.92074
susinidum_coc 1.84e-02 3.77e-02 0.24 0.62594
susinidum_pbc 4.83e-02 3.43e-02 1.98 0.15958
susinidum_mar 2.61e-02 3.22e-02 0.66 0.41722
psycom_dum_study 9.76e-02 1.09e-02 80.24 < 2e-16 ***
psycom_dum_with -2.87e-05 1.03e-02 0.00 0.99776
freq_cons_dum_5day 1.17e-03 9.13e-03 0.02 0.89794
cond_oc_dum_2inact -5.28e-02 1.32e-02 15.94 6.5e-05 ***
cond_oc_dum_3unemp -3.80e-02 9.94e-03 14.64 0.00013 ***
susprindum_coc 1.99e-02 1.57e-02 1.60 0.20597
susprindum_pbc 6.08e-02 1.31e-02 21.54 3.5e-06 ***
susprindum_mar 2.28e-02 2.21e-02 1.06 0.30356
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation structure = independence
Estimated Scale Parameters:
Estimate Std.err
(Intercept) 0.253 0.00356
Number of clusters: 14073 Maximum cluster size: 8 Warning in eval(family$initialize): non-integer #successes in a binomial glm!
| model | term | estimate | std.error | p.value | conf.low | conf.high |
|---|---|---|---|---|---|---|
| simplest, binomial | policonsumo2 | 1.10 | 0.04 | 0.0363 | 1.01 | 1.20 |
| full, binomial | policonsumo2 | 1.13 | 0.04 | 0.0067 | 1.03 | 1.23 |
| simplest, poisson | policonsumo2 | 1.03 | 0.01 | 0.0166 | 1.01 | 1.05 |
| full, poisson | policonsumo2 | 1.04 | 0.01 | 0.0036 | 1.01 | 1.06 |
Code
# can use to fit a weighted GEE
mw_alt2 <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_proc2_iiw_strata_alt_alt, family= binomial, weights=iiw_strata_st, corstr="independence")Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Code
#In eval(family$initialize) : non-integer #successes in a binomial glm!
#summary(mw)
mwpois_alt2 <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_proc2_iiw_strata_alt_alt, family= poisson(link = "log"), weights=iiw_strata_st, corstr="independence")
mwpois_alt2 <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_proc2_iiw_strata_alt_alt, family= poisson(link = "log"), weights=iiw_strata_st, corstr="independence")
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
mw_alt_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_proc2_iiw_strata_alt, family=binomial, weights=iiw_strata_st, corstr="independence") Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Code
#In eval(family$initialize) : non-integer #successes in a binomial glm!
mw_alt2_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_proc2_iiw_strata_alt_alt, family=binomial, weights=iiw_strata_st, corstr="independence")Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Code
mwpois_alt_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_proc2_iiw_strata_alt, family= poisson(link = "log"), weights=iiw_strata_st, corstr="independence")
mwpois_alt2_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_proc2_iiw_strata_alt_alt, family= poisson(link = "log"), weights=iiw_strata_st, corstr="independence")
invisible("Model fit")
#broom::glance(mwpois_full)
#QIC(mw)
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
rbind.data.frame(cbind.data.frame(model="simplest, binomial", broom::tidy(mw_alt, exponentiate = T, conf.int=T)),cbind.data.frame(model="full, binomial", broom::tidy(mw_alt_full, exponentiate = T, conf.int=T)),cbind.data.frame(model="simplest, poisson", broom::tidy(mwpois_alt, exponentiate = T, conf.int=T)),cbind.data.frame(model="full, poisson", broom::tidy(mwpois_alt_full, exponentiate = T, conf.int=T)),
cbind.data.frame(model="simplest, binomial (alt)", broom::tidy(mw_alt2, exponentiate = T, conf.int=T)),cbind.data.frame(model="full, binomial (alt)", broom::tidy(mw_alt2_full, exponentiate = T, conf.int=T)),cbind.data.frame(model="simplest, poisson (alt)", broom::tidy(mwpois_alt2, exponentiate = T, conf.int=T)),cbind.data.frame(model="full, poisson (alt)", broom::tidy(mwpois_alt2_full, exponentiate = T, conf.int=T))
) %>% dplyr::mutate_at(c("estimate","std.error","conf.low","conf.high"), ~sprintf("%1.2f",.))%>% dplyr::mutate_at(c("p.value"), ~sprintf("%1.4f",.)) %>% dplyr::select(-statistic) %>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/geew_strata_alt_240503.csv"), dec=",", sep="\t")
knitr::kable(dplyr::filter(.,term=="policonsumo2"), size=10, format="markdown", caption="GEE Models, alternative weighting (Stratifying)")
}| model | term | estimate | std.error | p.value | conf.low | conf.high |
|---|---|---|---|---|---|---|
| simplest, binomial | policonsumo2 | 1.10 | 0.05 | 0.0457 | 1.00 | 1.21 |
| full, binomial | policonsumo2 | 1.06 | 0.04 | 0.1576 | 0.98 | 1.14 |
| simplest, poisson | policonsumo2 | 1.03 | 0.01 | 0.0184 | 1.01 | 1.06 |
| full, poisson | policonsumo2 | 1.02 | 0.01 | 0.1014 | 1.00 | 1.04 |
| simplest, binomial (alt) | policonsumo2 | 1.03 | 0.05 | 0.4846 | 0.94 | 1.14 |
| full, binomial (alt) | policonsumo2 | 1.07 | 0.05 | 0.1612 | 0.97 | 1.18 |
| simplest, poisson (alt) | policonsumo2 | 1.01 | 0.01 | 0.3140 | 0.99 | 1.04 |
| full, poisson (alt) | policonsumo2 | 1.02 | 0.01 | 0.1078 | 1.00 | 1.05 |
No corrections
Code
#load("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/20240503_avance_iiw.RData")
iiw_nocorr<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1.lag+
comp_bpsc_y3_severe.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= data_mine_miss_restr_proc2 %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= maxfu_restr_df %>% as.data.frame(),
# formulanull = Surv(time.lag,time,event)~ edad_al_ing_1 +
# ano_nac_corr,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome", "comp_bpsc_y3_severe", "less_90d_tr1", "policonsumo2"),
lagfirst= c(2.95082,4.499811/2,0,0,0,0), #90/30.5 4.499811 es 90 días
first= T
)Warning in Surv(time.lag, time, event): Stop time must be > start time, NA created
Code
iiw_nocorr_alt<-
iiw.weights(Surv(time.lag,time,event)~
cluster(id)+
tr_outcome.lag +
log_dias_treat_imp_sin_na.lag +
less_90d_tr1.lag+
comp_bpsc_y3_severe.lag +
policonsumo2.lag +
edad_al_ing_1 +
ano_nac_corr +
susinidum_coc +
susinidum_oh +
susinidum_pbc +
susinidum_mar +
psycom_dum_with +
psycom_dum_study +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
data= data_mine_miss_restr_proc2 %>% as.data.frame(),
id= "id",
time= "time",
event= "event", #character string indicating which column of the data indicates whether or not a visit occurred. If every row corresponds to a visit, then this column will consist entirely of ones
maxfu= maxfu_restr_df %>% as.data.frame(),
# formulanull = Surv(time.lag,time,event)~ edad_al_ing_1 +
# ano_nac_corr,
invariant= c("edad_al_ing_1", "ano_nac_corr", "susinidum_coc", "susinidum_oh", "susinidum_pbc", "susinidum_mar", "psycom_dum_with", "psycom_dum_study", "freq_cons_dum_5day", "cond_oc_dum_2inact", "cond_oc_dum_3unemp", "susprindum_oh", "susprindum_coc", "susprindum_pbc", "susprindum_mar"),
lagvars= c("time", "log_dias_treat_imp_sin_na","tr_outcome", "comp_bpsc_y3_severe", "less_90d_tr1", "policonsumo2"),
lagfirst= c(2.95082,4.499811,1,1,1,1), #90/30.5 2.95082 ; 4.499811 es 90 días
first= T
)Warning in Surv(time.lag, time, event): Stop time must be > start time, NA created
Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_Code
summary(iiw_nocorr$iiw) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.21 1.00 1.62 1.84 2.56 5.99 Code
summary(iiw_nocorr_alt$iiw) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.1 0.4 0.7 0.8 1.0 102.3 Code
data_mine_miss_restr_proc2$iiw_nocorr<-iiw_nocorr$iiw
data_mine_miss_restr_proc2$iiw_nocorr_alt<-iiw_nocorr_alt$iiw
data_mine_miss_restr_proc2$iiw_nocorr_st<-data_mine_miss_restr_proc2$iiw_nocorr
data_mine_miss_restr_proc2$iiw_nocorr_st[data_mine_miss_restr_proc2$iiw_nocorr>quantile(data_mine_miss_restr_proc2$iiw_nocorr,0.975)] <- quantile(data_mine_miss_restr_proc2$iiw_nocorr,0.975)
data_mine_miss_restr_proc2$iiw_nocorr_st[data_mine_miss_restr_proc2_iiw_after_ph$iiw_nocorr<quantile(data_mine_miss_restr_proc2$iiw_nocorr,0.025)] <- quantile(data_mine_miss_restr_proc2$iiw_nocorr,0.025)
data_mine_miss_restr_proc2$iiw_nocorr_alt_st<-data_mine_miss_restr_proc2$iiw_nocorr_alt
data_mine_miss_restr_proc2$iiw_nocorr_alt_st[data_mine_miss_restr_proc2$iiw_nocorr_alt>quantile(data_mine_miss_restr_proc2$iiw_nocorr_alt,0.975)] <- quantile(data_mine_miss_restr_proc2$iiw_nocorr_alt,0.975)
data_mine_miss_restr_proc2$iiw_nocorr_st[data_mine_miss_restr_proc2_iiw_after_ph$iiw_iiw_nocorr_altnocorr<quantile(data_mine_miss_restr_proc2$iiw_nocorr_alt,0.025)] <- quantile(data_mine_miss_restr_proc2$iiw_nocorr_alt,0.025)
summary(iiw_nocorr$iiw) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.21 1.00 1.62 1.84 2.56 5.99 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.53 1.00 1.54 1.87 2.67 6.03
summary(iiw_nocorr_alt$iiw) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.1 0.4 0.7 0.8 1.0 102.3 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.1 0.4 0.7 0.8 1.0 102.3
summary(data_mine_miss_restr_proc2$iiw_nocorr) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.21 1.00 1.62 1.84 2.56 5.99 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.53 1.00 1.54 1.87 2.67 6.03
summary(data_mine_miss_restr_proc2$iiw_nocorr_st) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.21 1.00 1.62 1.83 2.56 3.79 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.53 1.00 1.54 1.86 2.67 3.89
summary(data_mine_miss_restr_proc2$iiw_nocorr_alt_st) Min. 1st Qu. Median Mean 3rd Qu. Max.
0.129 0.418 0.676 0.697 1.000 1.000 Code
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 0.129 0.418 0.676 0.697 1.000 1.000
rbind.data.frame(
cbind.data.frame(model= "Primary (no corrections, lag=0)", t(matrix(summary(data_mine_miss_restr_proc2$iiw_nocorr_st)))),
cbind.data.frame(model= "Alternative (no corrections, lag=1)", t(matrix(summary(data_mine_miss_restr_proc2$iiw_nocorr_alt_st)))))%>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/iiw_nocorr_240503.csv"), dec=",", sep="\t")
knitr::kable(., size=10, format="markdown", caption="Weights (No corrections)", col.names= c("Weight",attr(summary(data_mine_miss_restr_proc2$iiw_nocorr_st),"names")) )
}| Weight | Min. | 1st Qu. | Median | Mean | 3rd Qu. | Max. |
|---|---|---|---|---|---|---|
| Primary (no corrections, lag=0) | 0.214 | 1.000 | 1.617 | 1.834 | 2.56 | 3.79 |
| Alternative (no corrections, lag=1) | 0.129 | 0.418 | 0.676 | 0.697 | 1.00 | 1.00 |
GEE
Code
#The id= option is where we specify the clusters within which we have repeated observations that may be correlated, and tehe scale.fix=T option is to avoid R's default approach of introducing an overdispersion scale parameter.
#However, caution is advised when employing QIC and its use should not be routine, see Wang et al. (2015). growing use of quasi-likelihood-based information criteria for longitudinal data to select a working correlation structure in a generalized estimating equation framework. [https://onlinelibrary.wiley.com/doi/full/10.1002/sta4.95]
#poisson(link = "log"),
#https://www.thelancet.com/cms/10.1016/S2468-2667(22)00042-1/attachment/eb69ffc4-cb6c-4a47-9524-94e1ba9302b2/mmc1.pdf
#https://www.thelancet.com/cms/10.1016/S2468-2667(22)00201-8/attachment/05a17ae1-1e31-4ffd-a727-78ec915dc03b/mmc1.pdf
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
invisible("Must be independence structure")
m_nocorr <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_restr_proc2, family=binomial, corstr="independence")
summary(m_nocorr)
Call:
geeglm(formula = tr_outcome ~ policonsumo2 + edad_al_ing_1 +
ano_nac_corr + susinidum_oh + susinidum_coc + susinidum_pbc +
susinidum_mar + psycom_dum_study + psycom_dum_with + freq_cons_dum_5day +
cond_oc_dum_2inact + cond_oc_dum_3unemp + susprindum_coc +
susprindum_pbc + susprindum_mar, family = binomial, data = data_mine_miss_restr_proc2,
id = id, corstr = "independence")
Coefficients:
Estimate Std.err Wald Pr(>|W|)
(Intercept) -43.33863 13.16878 10.83 0.00100 ***
policonsumo2 0.05760 0.03597 2.56 0.10927
edad_al_ing_1 -0.00669 0.00668 1.01 0.31608
ano_nac_corr 0.02251 0.00654 11.83 0.00058 ***
susinidum_oh -0.07388 0.09434 0.61 0.43356
susinidum_coc -0.01507 0.11807 0.02 0.89845
susinidum_pbc 0.14542 0.10781 1.82 0.17737
susinidum_mar 0.04749 0.09658 0.24 0.62290
psycom_dum_study 0.86346 0.04632 347.54 < 2e-16 ***
psycom_dum_with 0.06046 0.03263 3.43 0.06392 .
freq_cons_dum_5day -0.02979 0.03089 0.93 0.33480
cond_oc_dum_2inact -0.25843 0.04290 36.29 1.7e-09 ***
cond_oc_dum_3unemp -0.19282 0.03449 31.25 2.3e-08 ***
susprindum_coc 0.02959 0.04840 0.37 0.54105
susprindum_pbc 0.23707 0.04089 33.61 6.7e-09 ***
susprindum_mar 0.06276 0.07496 0.70 0.40244
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation structure = independence
Estimated Scale Parameters:
Estimate Std.err
(Intercept) 1 0.0162
Number of clusters: 13317 Maximum cluster size: 10 Code
m_nocorr_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_restr_proc2, family=binomial, corstr="independence")
rbind.data.frame(cbind.data.frame(model="simplest", broom::tidy(m_nocorr, exponentiate = T, conf.int=T)),cbind.data.frame(model="full", broom::tidy(m_nocorr_full, exponentiate = T, conf.int=T))) %>% dplyr::mutate_at(c("estimate","std.error","conf.low","conf.high"), ~sprintf("%1.2f",.))%>% dplyr::mutate_at(c("p.value"), ~sprintf("%1.4f",.)) %>% dplyr::select(-statistic) %>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/gee_nocorr_240503.csv"), dec=",", sep="\t")
knitr::kable(dplyr::filter(.,term=="policonsumo2"), size=10, format="markdown", caption="GEE Models (no correction)")
}| model | term | estimate | std.error | p.value | conf.low | conf.high |
|---|---|---|---|---|---|---|
| simplest | policonsumo2 | 1.06 | 0.04 | 0.1093 | 0.99 | 1.14 |
| full | policonsumo2 | 1.10 | 0.04 | 0.0079 | 1.03 | 1.18 |
Code
data_mine_miss_restr_proc2$edad_al_ing_1_st = data_mine_miss_restr_proc2$edad_al_ing_1 - mean(data_mine_miss_restr_proc2$edad_al_ing_1)
# can use to fit a weighted GEE
mw_nocorr <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_restr_proc2, family= binomial, weights=iiw_nocorr_st, corstr="independence")Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Code
#In eval(family$initialize) : non-integer #successes in a binomial glm!
#summary(mw)
mw_nocorr_pois <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_restr_proc2, family= poisson(link = "log"), weights=iiw_nocorr_st, corstr="independence")
summary(mwpois)
Call:
geeglm(formula = tr_outcome ~ policonsumo2 + edad_al_ing_1 +
ano_nac_corr + susinidum_oh + susinidum_coc + susinidum_pbc +
susinidum_mar + psycom_dum_study + psycom_dum_with + freq_cons_dum_5day +
cond_oc_dum_2inact + cond_oc_dum_3unemp + susprindum_coc +
susprindum_pbc + susprindum_mar, family = poisson(link = "log"),
data = data_mine_miss_restr_proc2_iiw_after_ph, weights = iiw_after_ph_st,
id = id, corstr = "independence")
Coefficients:
Estimate Std.err Wald Pr(>|W|)
(Intercept) -6.48845 3.56281 3.32 0.0686 .
policonsumo2 0.02903 0.01098 6.99 0.0082 **
edad_al_ing_1 -0.00520 0.00182 8.13 0.0044 **
ano_nac_corr 0.00318 0.00177 3.23 0.0724 .
susinidum_oh -0.01208 0.02868 0.18 0.6736
susinidum_coc 0.00578 0.03446 0.03 0.8668
susinidum_pbc 0.03984 0.03091 1.66 0.1975
susinidum_mar 0.01794 0.02878 0.39 0.5331
psycom_dum_study 0.13012 0.01024 161.40 < 2e-16 ***
psycom_dum_with 0.00609 0.01002 0.37 0.5436
freq_cons_dum_5day 0.00429 0.00860 0.25 0.6179
cond_oc_dum_2inact -0.05450 0.01246 19.14 1.2e-05 ***
cond_oc_dum_3unemp -0.03855 0.00943 16.71 4.4e-05 ***
susprindum_coc 0.01730 0.01487 1.35 0.2447
susprindum_pbc 0.05929 0.01247 22.60 2.0e-06 ***
susprindum_mar 0.01361 0.02142 0.40 0.5252
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation structure = independence
Estimated Scale Parameters:
Estimate Std.err
(Intercept) 0.256 0.00339
Number of clusters: 13317 Maximum cluster size: 10 Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
mw_nocorr_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_restr_proc2, family=binomial, weights=iiw_nocorr_st, corstr="independence") Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Code
#In eval(family$initialize) : non-integer #successes in a binomial glm!
mw_nocorr_pois_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_restr_proc2, family= poisson(link = "log"), weights=iiw_nocorr_st, corstr="independence")
invisible("Model fit")
#broom::glance(mwpois_full)
#QIC(mw)
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
rbind.data.frame(cbind.data.frame(model="simplest, binomial", broom::tidy(mw_nocorr, exponentiate = T, conf.int=T)),cbind.data.frame(model="full, binomial", broom::tidy(mw_nocorr_full, exponentiate = T, conf.int=T)),cbind.data.frame(model="simplest, poisson", broom::tidy(mw_nocorr_pois, exponentiate = T, conf.int=T)),cbind.data.frame(model="full, poisson", broom::tidy(mw_nocorr_pois_full, exponentiate = T, conf.int=T))) %>% dplyr::mutate_at(c("estimate","std.error","conf.low","conf.high"), ~sprintf("%1.2f",.))%>% dplyr::mutate_at(c("p.value"), ~sprintf("%1.4f",.)) %>% dplyr::select(-statistic) %>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/geew_nocorr_240503.csv"), dec=",", sep="\t")
knitr::kable(dplyr::filter(.,term=="policonsumo2"), size=10, format="markdown", caption="GEE Models, weighted (no correction)")
}| model | term | estimate | std.error | p.value | conf.low | conf.high |
|---|---|---|---|---|---|---|
| simplest, binomial | policonsumo2 | 1.08 | 0.04 | 0.0548 | 1.00 | 1.16 |
| full, binomial | policonsumo2 | 1.11 | 0.04 | 0.0063 | 1.03 | 1.20 |
| simplest, poisson | policonsumo2 | 1.02 | 0.01 | 0.0215 | 1.00 | 1.04 |
| full, poisson | policonsumo2 | 1.03 | 0.01 | 0.0029 | 1.01 | 1.05 |
Code
# can use to fit a weighted GEE
mw_nocorr_alt <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_restr_proc2, family= binomial, weights=iiw_nocorr_alt_st, corstr="independence")Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Code
#In eval(family$initialize) : non-integer #successes in a binomial glm!
#summary(mw)
mw_nocorr_pois_alt <- geeglm( tr_outcome ~ policonsumo2 +
edad_al_ing_1 +
ano_nac_corr +
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar +
psycom_dum_study +
psycom_dum_with +
freq_cons_dum_5day +
cond_oc_dum_2inact +
cond_oc_dum_3unemp +
susprindum_coc +
susprindum_pbc +
susprindum_mar,
id=id, data=data_mine_miss_restr_proc2, family= poisson(link = "log"), weights=iiw_nocorr_alt_st, corstr="independence")
summary(mwpois)
Call:
geeglm(formula = tr_outcome ~ policonsumo2 + edad_al_ing_1 +
ano_nac_corr + susinidum_oh + susinidum_coc + susinidum_pbc +
susinidum_mar + psycom_dum_study + psycom_dum_with + freq_cons_dum_5day +
cond_oc_dum_2inact + cond_oc_dum_3unemp + susprindum_coc +
susprindum_pbc + susprindum_mar, family = poisson(link = "log"),
data = data_mine_miss_restr_proc2_iiw_after_ph, weights = iiw_after_ph_st,
id = id, corstr = "independence")
Coefficients:
Estimate Std.err Wald Pr(>|W|)
(Intercept) -6.48845 3.56281 3.32 0.0686 .
policonsumo2 0.02903 0.01098 6.99 0.0082 **
edad_al_ing_1 -0.00520 0.00182 8.13 0.0044 **
ano_nac_corr 0.00318 0.00177 3.23 0.0724 .
susinidum_oh -0.01208 0.02868 0.18 0.6736
susinidum_coc 0.00578 0.03446 0.03 0.8668
susinidum_pbc 0.03984 0.03091 1.66 0.1975
susinidum_mar 0.01794 0.02878 0.39 0.5331
psycom_dum_study 0.13012 0.01024 161.40 < 2e-16 ***
psycom_dum_with 0.00609 0.01002 0.37 0.5436
freq_cons_dum_5day 0.00429 0.00860 0.25 0.6179
cond_oc_dum_2inact -0.05450 0.01246 19.14 1.2e-05 ***
cond_oc_dum_3unemp -0.03855 0.00943 16.71 4.4e-05 ***
susprindum_coc 0.01730 0.01487 1.35 0.2447
susprindum_pbc 0.05929 0.01247 22.60 2.0e-06 ***
susprindum_mar 0.01361 0.02142 0.40 0.5252
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation structure = independence
Estimated Scale Parameters:
Estimate Std.err
(Intercept) 0.256 0.00339
Number of clusters: 13317 Maximum cluster size: 10 Code
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
mw_nocorr_alt_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_restr_proc2, family=binomial, weights=iiw_nocorr_alt_st, corstr="independence") Warning in eval(family$initialize): non-integer #successes in a binomial glm!
Code
#In eval(family$initialize) : non-integer #successes in a binomial glm!
mw_nocorr_pois_alt_full <- geeglm( tr_outcome ~ policonsumo2 +
comp_bpsc_y2_moderate+
comp_bpsc_y3_severe+
edad_al_ing_1+
ano_nac_corr+
esc_dum_rec_3prim+
esc_dum_rec_2high +
susprindum_oh +
susprindum_coc +
susprindum_pbc +
susprindum_mar+
freq_cons_dum_5day+
freq_cons_dum_44to6wk+
freq_cons_dum_32to3wk+
freq_cons_dum_21wkmore+
cond_oc_dum_3unemp+
cond_oc_dum_2inact+
viv_dum_illegal+
viv_dum_own+
viv_dum_rent+
viv_dum_temp+
psycom_dum_with+
psycom_dum_study+
susinidum_oh +
susinidum_coc +
susinidum_pbc +
susinidum_mar+
cohab_dum_alone+
cohab_dum_fam_or+
cohab_dum_cpl_child,
id=id, data=data_mine_miss_restr_proc2, family= poisson(link = "log"), weights=iiw_nocorr_alt_st, corstr="independence")
invisible("Model fit")
#broom::glance(mwpois_full)
#QIC(mw)
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_#_
rbind.data.frame(cbind.data.frame(model="simplest, binomial", broom::tidy(mw_nocorr_alt, exponentiate = T, conf.int=T)), cbind.data.frame(model="full, binomial", broom::tidy(mw_nocorr_alt_full, exponentiate = T, conf.int=T)), cbind.data.frame(model="simplest, poisson", broom::tidy(mw_nocorr_pois_alt, exponentiate = T, conf.int=T)),cbind.data.frame(model="full, poisson", broom::tidy(mw_nocorr_pois_alt_full, exponentiate = T, conf.int=T))) %>% dplyr::mutate_at(c("estimate","std.error","conf.low","conf.high"), ~sprintf("%1.2f",.))%>% dplyr::mutate_at(c("p.value"), ~sprintf("%1.4f",.)) %>% dplyr::select(-statistic) %>%
{
write.table(., file = paste0(getwd(),"/_proposal_grant/2023/geew_nocorr_alt_240503.csv"), dec=",", sep="\t")
knitr::kable(dplyr::filter(.,term=="policonsumo2"), size=10, format="markdown", caption="GEE Models, alternative weighting (no correction)")
}| model | term | estimate | std.error | p.value | conf.low | conf.high |
|---|---|---|---|---|---|---|
| simplest, binomial | policonsumo2 | 1.04 | 0.04 | 0.2923 | 0.97 | 1.12 |
| full, binomial | policonsumo2 | 1.09 | 0.04 | 0.0321 | 1.01 | 1.17 |
| simplest, poisson | policonsumo2 | 1.01 | 0.01 | 0.1631 | 0.99 | 1.03 |
| full, poisson | policonsumo2 | 1.02 | 0.01 | 0.0206 | 1.00 | 1.04 |
Session info
Code
message(paste0("R library: ", Sys.getenv("R_LIBS_USER")))Code
message(paste0("Date: ",withr::with_locale(new = c('LC_TIME' = 'C'), code =Sys.time())))Code
message(paste0("Editor context: ", path))Code
sesion_info <- devtools::session_info()
dplyr::select(
tibble::as_tibble(sesion_info$packages),
c(package, loadedversion, source)
) %>%
kable(caption = "R packages", format = "markdown",
col.names = c("Row number", "Package", "Version"),
row.names = FALSE,
align = c("c", "l", "r"))| Row number | Package | Version |
|---|---|---|
| abind | 1.4-5 | CRAN (R 4.1.1) |
| arrow | 15.0.1 | CRAN (R 4.1.2) |
| assertthat | 0.2.1 | CRAN (R 4.1.2) |
| backports | 1.4.1 | CRAN (R 4.1.2) |
| base64enc | 0.1-3 | CRAN (R 4.1.1) |
| BiocManager | 1.30.18 | CRAN (R 4.1.3) |
| bit | 4.0.4 | CRAN (R 4.1.2) |
| bit64 | 4.0.5 | CRAN (R 4.1.2) |
| blob | 1.2.3 | CRAN (R 4.1.3) |
| boot | 1.3-28.1 | CRAN (R 4.1.3) |
| broom | 1.0.1 | CRAN (R 4.1.3) |
| cachem | 1.0.6 | CRAN (R 4.1.2) |
| callr | 3.7.2 | CRAN (R 4.1.3) |
| car | 3.1-0 | CRAN (R 4.1.3) |
| carData | 3.0-5 | CRAN (R 4.1.2) |
| cellranger | 1.1.0 | CRAN (R 4.1.2) |
| checkmate | 2.1.0 | CRAN (R 4.1.3) |
| chron | 2.3-58 | CRAN (R 4.1.3) |
| class | 7.3-20 | CRAN (R 4.1.3) |
| cli | 3.4.1 | CRAN (R 4.1.3) |
| clisymbols | 1.2.0 | CRAN (R 4.1.3) |
| cluster | 2.1.4 | CRAN (R 4.1.3) |
| codetools | 0.2-19 | CRAN (R 4.1.3) |
| colorspace | 2.0-3 | CRAN (R 4.1.2) |
| compareGroups | 4.5.1 | CRAN (R 4.1.3) |
| cowplot | 1.1.1 | CRAN (R 4.1.2) |
| crayon | 1.5.2 | CRAN (R 4.1.3) |
| curl | 4.3.3 | CRAN (R 4.1.3) |
| DALEX | 2.4.3 | CRAN (R 4.1.3) |
| data.table | 1.14.2 | CRAN (R 4.1.2) |
| data.tree | 1.0.0 | CRAN (R 4.1.3) |
| DBI | 1.1.3 | CRAN (R 4.1.3) |
| dbplyr | 2.2.1 | CRAN (R 4.1.3) |
| deldir | 1.0-6 | CRAN (R 4.1.1) |
| devtools | 2.4.4 | CRAN (R 4.1.3) |
| DiagrammeR | 1.0.9 | CRAN (R 4.1.3) |
| DiagrammeRsvg | 0.1 | CRAN (R 4.1.3) |
| digest | 0.6.29 | CRAN (R 4.1.3) |
| doRNG | 1.8.2 | CRAN (R 4.1.3) |
| dplyr | 1.0.10 | CRAN (R 4.1.3) |
| DT | 0.26 | CRAN (R 4.1.3) |
| e1071 | 1.7-11 | CRAN (R 4.1.3) |
| easyalluvial | 0.3.2 | CRAN (R 4.1.2) |
| ellipsis | 0.3.2 | CRAN (R 4.1.2) |
| evaluate | 0.16 | CRAN (R 4.1.3) |
| fansi | 1.0.3 | CRAN (R 4.1.3) |
| farver | 2.1.1 | CRAN (R 4.1.3) |
| fastDummies | 1.7.3 | CRAN (R 4.1.2) |
| fastmap | 1.1.0 | CRAN (R 4.1.3) |
| flextable | 0.8.2 | CRAN (R 4.1.3) |
| FNN | 1.1.3.1 | CRAN (R 4.1.3) |
| forcats | 0.5.2 | CRAN (R 4.1.3) |
| foreach | 1.5.2 | CRAN (R 4.1.2) |
| foreign | 0.8-83 | CRAN (R 4.1.3) |
| Formula | 1.2-4 | CRAN (R 4.1.1) |
| fs | 1.5.2 | CRAN (R 4.1.3) |
| future | 1.28.0 | CRAN (R 4.1.3) |
| future.apply | 1.9.1 | CRAN (R 4.1.3) |
| gargle | 1.2.1 | CRAN (R 4.1.3) |
| gdtools | 0.2.4 | CRAN (R 4.1.2) |
| gee | 4.13-26 | CRAN (R 4.1.2) |
| geepack | 1.3.9 | CRAN (R 4.1.3) |
| geesmv | 1.3 | CRAN (R 4.1.3) |
| generics | 0.1.3 | CRAN (R 4.1.3) |
| ggalluvial | 0.12.5 | CRAN (R 4.1.3) |
| ggplot2 | 3.4.1 | CRAN (R 4.1.3) |
| ggpubr | 0.4.0 | CRAN (R 4.1.2) |
| ggridges | 0.5.4 | CRAN (R 4.1.3) |
| ggsignif | 0.6.4 | CRAN (R 4.1.3) |
| globals | 0.16.1 | CRAN (R 4.1.3) |
| glue | 1.6.2 | CRAN (R 4.1.3) |
| googledrive | 2.0.0 | CRAN (R 4.1.2) |
| googlesheets4 | 1.0.1 | CRAN (R 4.1.3) |
| gower | 1.0.0 | CRAN (R 4.1.2) |
| gridExtra | 2.3 | CRAN (R 4.1.2) |
| gsubfn | 0.7 | CRAN (R 4.1.2) |
| gtable | 0.3.1 | CRAN (R 4.1.3) |
| hardhat | 1.2.0 | CRAN (R 4.1.3) |
| HardyWeinberg | 1.7.5 | CRAN (R 4.1.3) |
| haven | 2.5.1 | CRAN (R 4.1.3) |
| highr | 0.9 | CRAN (R 4.1.3) |
| Hmisc | 4.7-1 | CRAN (R 4.1.3) |
| hms | 1.1.2 | CRAN (R 4.1.3) |
| htmlTable | 2.4.1 | CRAN (R 4.1.3) |
| htmltools | 0.5.3 | CRAN (R 4.1.3) |
| htmlwidgets | 1.5.4 | CRAN (R 4.1.2) |
| httpuv | 1.6.6 | CRAN (R 4.1.3) |
| httr | 1.4.4 | CRAN (R 4.1.3) |
| interp | 1.1-3 | CRAN (R 4.1.3) |
| ipred | 0.9-13 | CRAN (R 4.1.3) |
| IrregLong | 0.3.4 | CRAN (R 4.1.3) |
| iterators | 1.0.14 | CRAN (R 4.1.2) |
| itertools | 0.1-3 | CRAN (R 4.1.3) |
| janitor | 2.1.0 | CRAN (R 4.1.2) |
| jpeg | 0.1-9 | CRAN (R 4.1.1) |
| jsonlite | 1.8.2 | CRAN (R 4.1.3) |
| jtools | 2.2.0 | CRAN (R 4.1.3) |
| kableExtra | 1.3.4 | CRAN (R 4.1.3) |
| km.ci | 0.5-6 | CRAN (R 4.1.3) |
| KMsurv | 0.1-5 | CRAN (R 4.1.1) |
| knitr | 1.40 | CRAN (R 4.1.3) |
| labeling | 0.4.2 | CRAN (R 4.1.1) |
| labelled | 2.10.0 | CRAN (R 4.1.3) |
| later | 1.3.0 | CRAN (R 4.1.2) |
| lattice | 0.20-45 | CRAN (R 4.1.1) |
| latticeExtra | 0.6-30 | CRAN (R 4.1.3) |
| lava | 1.6.10 | CRAN (R 4.1.2) |
| lazyeval | 0.2.2 | CRAN (R 4.1.2) |
| lifecycle | 1.0.3 | CRAN (R 4.1.3) |
| listenv | 0.8.0 | CRAN (R 4.1.2) |
| lme4 | 1.1-30 | CRAN (R 4.1.3) |
| lubridate | 1.8.0 | CRAN (R 4.1.2) |
| magrittr | 2.0.3 | CRAN (R 4.1.3) |
| MASS | 7.3-58.1 | CRAN (R 4.1.3) |
| Matrix | 1.5-1 | CRAN (R 4.1.3) |
| matrixcalc | 1.0-6 | CRAN (R 4.1.3) |
| MatrixModels | 0.5-1 | CRAN (R 4.1.3) |
| memoise | 2.0.1 | CRAN (R 4.1.2) |
| MEMSS | 0.9-3 | CRAN (R 4.1.3) |
| mice | 3.14.0 | CRAN (R 4.1.2) |
| mime | 0.12 | CRAN (R 4.1.1) |
| miniUI | 0.1.1.1 | CRAN (R 4.1.3) |
| minqa | 1.2.4 | CRAN (R 4.1.2) |
| missForest | 1.5 | CRAN (R 4.1.3) |
| missRanger | 2.1.3 | CRAN (R 4.1.3) |
| mitools | 2.4 | CRAN (R 4.1.2) |
| mnormt | 2.1.0 | CRAN (R 4.1.3) |
| modelr | 0.1.9 | CRAN (R 4.1.3) |
| mstate | 0.3.2 | CRAN (R 4.1.2) |
| multcomp | 1.4-20 | CRAN (R 4.1.3) |
| munsell | 0.5.0 | CRAN (R 4.1.2) |
| mvtnorm | 1.1-3 | CRAN (R 4.1.1) |
| nlme | 3.1-159 | CRAN (R 4.1.3) |
| nloptr | 2.0.3 | CRAN (R 4.1.3) |
| nnet | 7.3-18 | CRAN (R 4.1.3) |
| officer | 0.4.4 | CRAN (R 4.1.3) |
| openxlsx | 4.2.5 | CRAN (R 4.1.2) |
| pacman | 0.5.1 | CRAN (R 4.1.3) |
| pander | 0.6.5 | CRAN (R 4.1.3) |
| parallelly | 1.32.1 | CRAN (R 4.1.3) |
| parcats | 0.0.5 | CRAN (R 4.1.2) |
| patchwork | 1.1.2 | CRAN (R 4.1.3) |
| pillar | 1.8.1 | CRAN (R 4.1.3) |
| pkgbuild | 1.3.1 | CRAN (R 4.1.2) |
| pkgconfig | 2.0.3 | CRAN (R 4.1.2) |
| pkgload | 1.3.0 | CRAN (R 4.1.3) |
| plotly | 4.10.0 | CRAN (R 4.1.2) |
| plyr | 1.8.7 | CRAN (R 4.1.3) |
| png | 0.1-7 | CRAN (R 4.1.1) |
| polspline | 1.1.20 | CRAN (R 4.1.3) |
| prettyunits | 1.1.1 | CRAN (R 4.1.2) |
| processx | 3.7.0 | CRAN (R 4.1.3) |
| prodlim | 2019.11.13 | CRAN (R 4.1.2) |
| profvis | 0.3.7 | CRAN (R 4.1.3) |
| progressr | 0.11.0 | CRAN (R 4.1.3) |
| promises | 1.2.0.1 | CRAN (R 4.1.2) |
| proto | 1.0.0 | CRAN (R 4.1.2) |
| proxy | 0.4-27 | CRAN (R 4.1.3) |
| ps | 1.7.1 | CRAN (R 4.1.3) |
| psych | 2.2.9 | CRAN (R 4.1.3) |
| purrr | 0.3.5 | CRAN (R 4.1.3) |
| quantreg | 5.94 | CRAN (R 4.1.3) |
| R6 | 2.5.1 | CRAN (R 4.1.3) |
| ragg | 1.2.3 | CRAN (R 4.1.3) |
| randomForest | 4.7-1.1 | CRAN (R 4.1.3) |
| ranger | 0.14.1 | CRAN (R 4.1.3) |
| RColorBrewer | 1.1-3 | CRAN (R 4.1.3) |
| Rcpp | 1.0.9 | CRAN (R 4.1.3) |
| readr | 2.1.3 | CRAN (R 4.1.3) |
| readxl | 1.4.1 | CRAN (R 4.1.3) |
| recipes | 1.0.1 | CRAN (R 4.1.3) |
| remotes | 2.4.2 | CRAN (R 4.1.3) |
| renv | 1.0.1 | CRAN (R 4.1.2) |
| reprex | 2.0.2 | CRAN (R 4.1.3) |
| reshape2 | 1.4.4 | CRAN (R 4.1.2) |
| reticulate | 1.26 | CRAN (R 4.1.3) |
| rio | 0.5.29 | CRAN (R 4.1.2) |
| rlang | 1.0.6 | CRAN (R 4.1.3) |
| rmarkdown | 2.16 | CRAN (R 4.1.3) |
| rms | 6.3-0 | CRAN (R 4.1.3) |
| rngtools | 1.5.2 | CRAN (R 4.1.3) |
| rpart | 4.1.16 | CRAN (R 4.1.3) |
| Rsolnp | 1.16 | CRAN (R 4.1.3) |
| RSQLite | 2.2.18 | CRAN (R 4.1.3) |
| rstatix | 0.7.0 | CRAN (R 4.1.2) |
| rstudioapi | 0.14 | CRAN (R 4.1.3) |
| rvest | 1.0.3 | CRAN (R 4.1.3) |
| sandwich | 3.0-2 | CRAN (R 4.1.3) |
| scales | 1.2.1 | CRAN (R 4.1.3) |
| sessioninfo | 1.2.2 | CRAN (R 4.1.2) |
| shiny | 1.7.2 | CRAN (R 4.1.3) |
| snakecase | 0.11.0 | CRAN (R 4.1.2) |
| SparseM | 1.81 | CRAN (R 4.1.1) |
| sqldf | 0.4-11 | CRAN (R 4.1.3) |
| stringi | 1.7.6 | CRAN (R 4.1.2) |
| stringr | 1.4.1 | CRAN (R 4.1.3) |
| survex | 0.2.2 | CRAN (R 4.1.2) |
| survey | 4.1-1 | CRAN (R 4.1.2) |
| survival | 3.4-0 | CRAN (R 4.1.3) |
| survminer | 0.4.9 | CRAN (R 4.1.3) |
| survMisc | 0.5.6 | CRAN (R 4.1.3) |
| svglite | 2.1.0 | CRAN (R 4.1.2) |
| systemfonts | 1.0.4 | CRAN (R 4.1.2) |
| tableone | 0.13.2 | CRAN (R 4.1.3) |
| textshaping | 0.3.6 | CRAN (R 4.1.3) |
| TH.data | 1.1-1 | CRAN (R 4.1.3) |
| tibble | 3.1.8 | CRAN (R 4.1.3) |
| tidylog | 1.0.2 | CRAN (R 4.1.3) |
| tidyr | 1.2.1 | CRAN (R 4.1.3) |
| tidyselect | 1.2.0 | CRAN (R 4.1.3) |
| tidyverse | 1.3.2 | CRAN (R 4.1.3) |
| timeDate | 4021.106 | CRAN (R 4.1.3) |
| truncnorm | 1.0-8 | CRAN (R 4.1.3) |
| tzdb | 0.3.0 | CRAN (R 4.1.3) |
| urlchecker | 1.0.1 | CRAN (R 4.1.3) |
| usethis | 2.1.6 | CRAN (R 4.1.3) |
| utf8 | 1.2.2 | CRAN (R 4.1.2) |
| uuid | 1.1-0 | CRAN (R 4.1.3) |
| V8 | 4.2.1 | CRAN (R 4.1.3) |
| vctrs | 0.5.2 | CRAN (R 4.1.3) |
| viridisLite | 0.4.1 | CRAN (R 4.1.3) |
| visNetwork | 2.1.2 | CRAN (R 4.1.3) |
| webshot | 0.5.4 | CRAN (R 4.1.3) |
| withr | 2.5.0 | CRAN (R 4.1.2) |
| writexl | 1.4.0 | CRAN (R 4.1.3) |
| xfun | 0.33 | CRAN (R 4.1.3) |
| xml2 | 1.3.3 | CRAN (R 4.1.2) |
| xtable | 1.8-4 | CRAN (R 4.1.2) |
| yaml | 2.3.6 | CRAN (R 4.1.3) |
| zip | 2.2.1 | CRAN (R 4.1.3) |
| zoo | 1.8-11 | CRAN (R 4.1.3) |
Code
reticulate::py_list_packages()%>%
kable(caption = "Python packages", format = "markdown",
col.names = c("Package", "Version", "Requirement"),
row.names = FALSE,
align = c("c", "l", "r", "r"))| Package | Version | Requirement |
|---|---|---|
| asttokens | 2.4.1 | asttokens==2.4.1 |
| autograd | 1.6.2 | autograd==1.6.2 |
| autograd-gamma | 0.5.0 | autograd-gamma==0.5.0 |
| beautifulsoup4 | 4.12.3 | beautifulsoup4==4.12.3 |
| certifi | 2023.11.17 | certifi==2023.11.17 |
| cffi | 1.16.0 | cffi==1.16.0 |
| charset-normalizer | 3.3.2 | charset-normalizer==3.3.2 |
| click | 8.1.7 | click==8.1.7 |
| cloudpickle | 3.0.0 | cloudpickle==3.0.0 |
| colorama | 0.4.6 | colorama==0.4.6 |
| comm | 0.2.1 | comm==0.2.1 |
| contourpy | 1.2.0 | contourpy==1.2.0 |
| cycler | 0.12.1 | cycler==0.12.1 |
| debugpy | 1.8.0 | debugpy==1.8.0 |
| decorator | 5.1.1 | decorator==5.1.1 |
| delete-chrome-history-py | 0.1.8 | delete-chrome-history-py==0.1.8 |
| ecos | 2.0.13 | ecos==2.0.13 |
| executing | 2.0.1 | executing==2.0.1 |
| ffmpeg | 1.4 | ffmpeg==1.4 |
| ffmpeg-python | 0.2.0 | ffmpeg-python==0.2.0 |
| filedir | 0.0.3 | filedir==0.0.3 |
| filelock | 3.13.1 | filelock==3.13.1 |
| fonttools | 4.47.2 | fonttools==4.47.2 |
| formulaic | 1.0.1 | formulaic==1.0.1 |
| fsspec | 2023.12.2 | fsspec==2023.12.2 |
| future | 0.18.3 | future==0.18.3 |
| gTTS | 2.5.1 | gTTS==2.5.1 |
| idna | 3.6 | idna==3.6 |
| interface-meta | 1.3.0 | interface-meta==1.3.0 |
| ipykernel | 6.29.0 | ipykernel==6.29.0 |
| ipython | 8.20.0 | ipython==8.20.0 |
| jedi | 0.19.1 | jedi==0.19.1 |
| Jinja2 | 3.1.3 | Jinja2==3.1.3 |
| joblib | 1.4.0 | joblib==1.4.0 |
| jupyter_client | 8.6.0 | jupyter_client==8.6.0 |
| jupyter_core | 5.7.1 | jupyter_core==5.7.1 |
| kiwisolver | 1.4.5 | kiwisolver==1.4.5 |
| lifelines | 0.28.0 | lifelines==0.28.0 |
| llvmlite | 0.41.1 | llvmlite==0.41.1 |
| MarkupSafe | 2.1.4 | MarkupSafe==2.1.4 |
| matplotlib | 3.8.2 | matplotlib==3.8.2 |
| matplotlib-inline | 0.1.6 | matplotlib-inline==0.1.6 |
| more-itertools | 10.2.0 | more-itertools==10.2.0 |
| mpmath | 1.3.0 | mpmath==1.3.0 |
| multipledispatch | 1.0.0 | multipledispatch==1.0.0 |
| natsort | 8.4.0 | natsort==8.4.0 |
| nest-asyncio | 1.5.9 | nest-asyncio==1.5.9 |
| networkx | 3.2.1 | networkx==3.2.1 |
| numba | 0.58.1 | numba==0.58.1 |
| numexpr | 2.10.0 | numexpr==2.10.0 |
| numpy | 1.26.3 | numpy==1.26.3 |
| openai-whisper | 20231117 | openai-whisper==20231117 |
| osqp | 0.6.5 | osqp==0.6.5 |
| packaging | 23.2 | packaging==23.2 |
| pandas | 2.2.0 | pandas==2.2.0 |
| pandas-flavor | 0.6.0 | pandas-flavor==0.6.0 |
| parso | 0.8.3 | parso==0.8.3 |
| patsy | 0.5.6 | patsy==0.5.6 |
| pillow | 10.2.0 | pillow==10.2.0 |
| platformdirs | 4.1.0 | platformdirs==4.1.0 |
| prompt-toolkit | 3.0.43 | prompt-toolkit==3.0.43 |
| psutil | 5.9.8 | psutil==5.9.8 |
| pure-eval | 0.2.2 | pure-eval==0.2.2 |
| pyarrow | 15.0.0 | pyarrow==15.0.0 |
| pycparser | 2.22 | pycparser==2.22 |
| pydub | 0.25.1 | pydub==0.25.1 |
| Pygments | 2.17.2 | Pygments==2.17.2 |
| pyjanitor | 0.26.0 | pyjanitor==0.26.0 |
| pyparsing | 3.1.1 | pyparsing==3.1.1 |
| pyreadr | 0.5.0 | pyreadr==0.5.0 |
| python-dateutil | 2.8.2 | python-dateutil==2.8.2 |
| pytz | 2023.3.post1 | pytz==2023.3.post1 |
| pywin32 | 306 | pywin32==306 |
| pyzmq | 25.1.2 | pyzmq==25.1.2 |
| qdldl | 0.1.7.post1 | qdldl==0.1.7.post1 |
| regex | 2023.12.25 | regex==2023.12.25 |
| requests | 2.31.0 | requests==2.31.0 |
| rpy2 | 3.5.16 | rpy2==3.5.16 |
| scikit-learn | 1.3.2 | scikit-learn==1.3.2 |
| scikit-survival | 0.22.2 | scikit-survival==0.22.2 |
| scipy | 1.11.4 | scipy==1.11.4 |
| semantic-version | 2.10.0 | semantic-version==2.10.0 |
| setuptools-rust | 1.8.1 | setuptools-rust==1.8.1 |
| six | 1.16.0 | six==1.16.0 |
| soupsieve | 2.5 | soupsieve==2.5 |
| SpeechRecognition | 3.10.1 | SpeechRecognition==3.10.1 |
| spyder-kernels | 2.5.0 | spyder-kernels==2.5.0 |
| stack-data | 0.6.3 | stack-data==0.6.3 |
| statsmodels | 0.14.1 | statsmodels==0.14.1 |
| sympy | 1.12 | sympy==1.12 |
| target | 0.0.11 | target==0.0.11 |
| threadpoolctl | 3.4.0 | threadpoolctl==3.4.0 |
| tiktoken | 0.5.2 | tiktoken==0.5.2 |
| torch | 2.1.2 | torch==2.1.2 |
| tornado | 6.4 | tornado==6.4 |
| tqdm | 4.66.1 | tqdm==4.66.1 |
| traitlets | 5.14.1 | traitlets==5.14.1 |
| translator | 0.0.9 | translator==0.0.9 |
| typing_extensions | 4.9.0 | typing_extensions==4.9.0 |
| tzdata | 2023.4 | tzdata==2023.4 |
| tzlocal | 5.2 | tzlocal==5.2 |
| urllib3 | 2.1.0 | urllib3==2.1.0 |
| watchdog | 3.0.0 | watchdog==3.0.0 |
| wcwidth | 0.2.13 | wcwidth==0.2.13 |
| whisper | 1.1.10 | whisper==1.1.10 |
| wrapt | 1.16.0 | wrapt==1.16.0 |
| xarray | 2024.1.1 | xarray==2024.1.1 |
Save
Code
folder_path <- ifelse(dir.exists("E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/"),
"E:/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/",
"C:/Users/CISS Fondecyt/Mi unidad/Alvacast/SISTRAT 2022 (github)/_proposal_grant/2023/")
save.image(paste0(folder_path,"an_grant_23_24_3.RData"))